My First W/C system & some questions

[CrustyDemon]

Well after doing a lot of research, asking questions in forums and slowly purchasing parts yesterday day I finished my first w/c system.

With posting in this forum and other forums about setting up a dual loop system I've finally decided against the idea and looked into a more simplistic set up.

I originally purchased a Coolermaster Stacker 830 then decided it wasn't quite what I was after so I also purchased a Lian Li V2000B+ (fantastic case I must add).

Well, with my changes to the 2 loop set up and having already purchased the Stacker as well as 2 pumps I decided to w/c my original computer (as a bit of a learning curve) until I purchase everything for the Lian Li (7900GTX vid cards & intels new cpu when available etc). I'll finish the Lian Li later on this year.

OK... back to my w/c set up.

Specs
Case: Coolermaster Stacker 830
Mother: Abit IC7-G
CPU: Intel P4 3.4G Prescott
Memory: 2x Corsair TwinX1024-3200CT PRO 512MB DDR400 - 4 Sticks
Video: nVidia 6800GT
PSU: Antec True Blue 480W
H/Drive: 74G Raptor & 36G Raptor
Sound Card: Creative Fatal1ty X-FI
Controller: Coolermaster Aerogate 3
Drive: Pioneer DVD-R/RW 110 DBK
PCI Extra: Netgear Wireless adapter (for when I build next comp I'll move to another room)
Monitors: 2x Samsung 930BF 4ms 19" LCD's
Keyboard: Logitech G15
Mouse: Razer Copperhead
Speakers: Creative Gigaworks 7.1
Router/Modem: Belkin F1PI241EGau

Cooling
Radiator: Thermochill PA 120.2
Pump: Swiftech MCP655 12V
Reservior: Swiftech MCRES-MICRO
CPU Block: Swiftech Storm G4
GPU Block: Swiftech MCW60
Chip: Swiftech MCW30 (I wanted to keep noise down as the air cooler is noise as hell)
Tubing: Tygon 7/16" ID
Rad Fans: Tricod Science 12V/.12A
Coolant: Distilled water with Swiftech HydrX Coolant
Clamps: Stainless Steel worm type
Res Support: Sunbeam Wherever PCI Rack

Pics:

This is the only modification I had to do with this case. Drill two clearance holes in motherboard tray to attach pump.

Leak test. I decided to go Rad>GPU>Chip>CPU>Res>Pump>Rad due to a using least amount of tubing and I had read it doesn't make a lot of difference in temps which order they are.

I need to make some brackets to fix hard drives in place or come up with a better way.

I had shorter tubing coming from GPU to Chip, Chip to CPU & CPU to Res. I had to lengthen them as they were closing/kinking.

I may take out mesh and relace with an acrylic window.. not sure at this stage as this doesn't look too bad and quite transparent.

I may need ot fill in the gaps on each side of the rad. The fixing of the rad worked out quite good. I attached 5mm rubber to bottom of rad and the side bay holes lined up perfectly with the shroud holes in the PA120.2, no drilling necessary. I used 4 off 30mmx6G Pan head screws to attach.


Comments:
Noise - The system is quite silent but the things that make most noise is the memory, Hard drives & most of all the Antec power supply. (I have purchased a Seasonic S12 600w for my next system but didn't want to waste it on my current set up).

Overclock:
I have kept the o/c the same as when it was air cooled so I can compare temps.
CPU - Stock: 3400MHz - O/C: 3706MHz
GPU - Stock: Core 350MHz, Mem 1.0GHz - O/C: Core 410MHz, Mem 1.11GHz
DRAM Ratio: 1:1
CPU Coret: 1.3875v
DDR SDRAM: 2.7v
AGP: 1.55v

Temp - I'm a bit dissapointed with temps as I don't find them much different than my previous air cooled set up. Unless theres something wrong it's not that great. Just by using MM5 (i know it's not the best but it's how I guaged temps with air so as a direct comparison) - CPU Idle 42 - 42C, Load 52-55C case temp being 28.5C and room temp being 22C.

Questions:
A) I can fit a shround to rad on the inside of case - will this make much difference?
B ) I can fit 38mm or 25mm 2x fans to front or rad for a push/pull set up as well a shroud on inside will this be a waste of time/money & unnecessary noise or will it make temps a little better?
C) I have placed 4 off temp probes:
- 1. On the top of storm block as close to cpu as possible with block pre-fitted. I have it attached to a small part of the copper.
- 2. Attached to tubing at outlet of radiator.
- 3. Top of 74g Raptor.
- 4. Attached to case frame
Now with these placement of probes I'm not sure if I have them in the best spot/s? Getting close to the CPU is difficult.
D) Anyone have any suggestion as how I can better place or secure my two hard drives. I have considered getting a 150g Rapter and dropping down to one so not so cluttered.

Thanks in advance to any helpers.

 

[DFI Daishi]

a) there is little to no benifit to adding a shroud to the thermochill PA series of rads. they were made to be un restrictive enough to run optimally without a seperate shroud.

b) you'd do better running faster/more powerfull fans on one side than running push-pull on both sides. going push-pull doubles your noise, but only ups the airflow 1.5 times. that being said, what you have on there should be enough, depending on the speed that you run them at.

c) to monitor CPU temp, i would thermal epoxy the temp probe onto the side of the IHS. to monitor coolant temps, i would either waterproof a temp sensor and put it into the res or scratch a bit of paint off of one of the tanks on you rad and thermal epoxy the probe there.

d) no idea.

i'm supprised that your temps are as high as they are. i seem to remember that those boards tend to report high, but that only counts for so much. given the way that you've got your rad oriented, have you moved it around to make sure that there is no air trapped inside of it? rads tend to trap air unless the barbs are at the top.

all in all, i think that's a really nice build you have there.

 

[CrustyDemon]

Thanks DFI Daishi.

Just to confirm on the last point - Having the barbs at the top in vertical rad positions is better than having the barbs at bottom like i have it now? I wasn't sure which way was best to place rad barbs just ended that way for no reason, didn't think it mattered much.

 

[Dark Prodigy]

First thing... Very nice build and parts you have there..

I just finished my own W/C setup and asked some very similiar questions in a post I made today..check it out to get my temps and situation. I have been tinkering with my setup and wanted a lower noise, lower temp situation myself. Depending on how much noise you are wanting to deal with ( I mainly w/c for low noise) 2 higher performing fans in the same configuration you have will drop your temps nicely...

I'm currently experimenting with the push-pull method because I only have a single 120mm rad... Because I'd rather have low noise than super low temps, I'm going for 2 low noise/low rpm fans in a push-pull...that should lower my noise level significantly while maintaining around the same temps...My Sunon fan spins @ 3600 Rpm but but moves air @ 107 Cfm...it sounds like an overhead vent for a stove...

Your temps are a bit high..back up your monitoring with nTune and see what you get..

Hope this helps...

 

[DFI Daishi]

Thanks DFI Daishi.

Just to confirm on the last point - Having the barbs at the top in vertical rad positions is better than having the barbs at bottom like i have it now? I wasn't sure which way was best to place rad barbs just ended that way for no reason, didn't think it mattered much.

the orientation of the rad does not matter much, generally speaking. the only time it makes a difference is when you're trying to get all of the air out of your sysem.

i'm suggesting that you try tilting your rad or your whole case to see if you have some air stuck in your rad, which might account for unexpectedly high temps.

 

[Erasmus354]

A few things....first of all, how can your memory make noise?!?

Secondly, Daishi you are wrong about the shroud. The shroud has nothing to do with how restrictive the radiator is. The shroud is to minimize the dead spot from the fan motor. The PA series already has a pretty deep plenum area, but a bit of a shroud wouldn't hurt probably. And not all PA series rads are like the 120.x series. The PA 160 comes with a shroud stock.

 

[DFI Daishi]

Secondly, Daishi you are wrong about the shroud. The shroud has nothing to do with how restrictive the radiator is. The shroud is to minimize the dead spot from the fan motor. The PA series already has a pretty deep plenum area, but a bit of a shroud wouldn't hurt probably. And not all PA series rads are like the 120.x series. The PA 160 comes with a shroud stock.

first off, although i'm just going by memory here, i seem to remember marci saying that with the current PA series rads: push, pull, extra shroud or stock attachment all perform the same.

as for the dead spot, as i understand it: less restrictive rad, less pressure gradient across the rad, virtually no radial pressure gradient, virtually nothing to drive the formation of a dead spot.

 

[thewhiteguy]

the things that make most noise is the memory, Hard drives & most of all the Antec power supply

You have some sort of mechanical RAM?

 

[Imitation]

You might check the dial on the pump to make sure its running at full or at least half speed. if you aren't pushing much water with the pump on low, your temps will be much higher. But if its already on high then I don't have any other ideas really.

 

[Erasmus354]

first off, although i'm just going by memory here, i seem to remember marci saying that with the current PA series rads: push, pull, extra shroud or stock attachment all perform the same.

as for the dead spot, as i understand it: less restrictive rad, less pressure gradient across the rad, virtually no radial pressure gradient, virtually nothing to drive the formation of a dead spot.

The formation of the dead spot is simply because where the motor on the fan is you dont have air moving. It takes a certain depth for the air pushed by the fan to spread out to cover the dead spot from the fan motor. The dead spot is the same on any radiator no matter the restriction. It is governed by the fan and how far away from the radiator it is.

 

[annaconda]

Nice work, but your temperatures are little high. How much did you spend on that water cooling total includes shipping and tax?

 

[CrustyDemon]

I don't think the temp are reading correctly. As I'm told that using mb tems with water as opposed to air you loose the fan or fans blowing directly around the socket area which effects temp readings. The only way to do a direct comparison is to plug in a fan direct it at socket then measure temps.

I put a probe on the tube going into rad it read 35c also the one coming out of the rad was 41c so that wierd. Don't know how accuate these temp probes are....


Anyway.... I've decided to:

1. Put the pump on the case floor maybe this will help flow and be better for the pump longevity (so i'm told). The manufacurer does recommend having attached to case floor anyway.
2. Reseat the Storm just incase not seated correctly also I'll check holes incase some metal shavings have blocked some holes.
3. Put two SilenX LED fans on the front of rad to be controlled by aerogate mostly for the bling bling, it won't do much for the rad, maybe a little, I'll try to keep noise down. I'll change the tricod fans on rad back to Panaflo's (mediums) and put them on a shroud also controlled by aerogate. This will help air flow a little and cut down noise from Panaflow. Got 60mm to play in case front so may as well use it up.
4. Make some sort of HD bracket and fix hard HD's where pump is now so any heat will rise to top fan blow hole.
5. I'll change the loop direction and give the storm the first coolest liquid from the rad. I know it doesn't matter much which direction the blocks are in but going direct from rad to storm will help a little.
6. I'll check/shake the rad to see if there is any air in it... cheers DFI Daishi
7. I'll check pump dial as well...cheers Imitation

Keeping in mind I am using an o/c Prescott which gives off a lot of heat for a cpu.

My memory does make some noise maybe it has something to do with the LED display on it. You can easily tell noise coming from it matches the LED movements in sync. It's only when the memory is working hard. But it definately make some noise.

I'ts not cheap annaconda

 

[thewhiteguy]

6 degree rise at the block is way too high. That tells me your flow rate is really bad. Having your pump as the highest point in the loop is definitely a bad idea. Move the reservoir up higher than the pump or at least fill it up as high as you can.

You coolant looks pretty thick, is that the case?

Edit: with those tight bends, are you sure the force of the tubing isn't tilting the block, so it isn't flat on the CPU?

 

[CrustyDemon]

6 degree rise at the block is way too high.[/QUOTE
Read post above.

That tells me your flow rate is really bad.

Flow is quite good from what i know.

Having your pump as the highest point in the loop is definitely a bad idea. Move the reservoir up higher than the pump or at least fill it up as high as you can.

Read post above.

You coolant looks pretty thick, is that the case?

No... coolant is very close to distilled water viscosity. Coolant was mixed as per Swiftech directions.

Edit: with those tight bends, are you sure the force of the tubing isn't tilting the block, so it isn't flat on the CPU?

Read text under picture No 12. I lengethen the tubing so there was no kinks/bends in tube. AAlso, there isn't any unnecessary pressure on storm it's seated parallel to cpu althougth i will check a re-seat. Using 7/16" ID tygon tubing is quite flexible. I wanted to keep tubing lengths down to a minimum to increase flow.

 

[Erasmus354]

6 degree rise at the block is way too high. That tells me your flow rate is really bad. Having your pump as the highest point in the loop is definitely a bad idea. Move the reservoir up higher than the pump or at least fill it up as high as you can.

A large delta between idle and load temperatures is usually more indicative of an undersized radiator than poor flow rates.

 

[DangerIsGo]

You have some sort of mechanical RAM?

Hahahah. I was just about to say..RAM making noise? then you gotta have SOMETHING wrong

 

[CrustyDemon]

ahhh...*sigh*

 

[CrustyDemon]

Set up take 2:

I changed a few things around.

1. Put a shroud on PA120.2 and changed fans to push instead of pull. Also, put in some Panaflo M fans in but this doesn't change much from Tricods. Just give me the option to turn them up in really hot days.
2. Flipped rad to have barbs at top. Better to eliminate any air pockts.
3. Pulled Storm apart incase of any blockages from build shavings etc. Didn't see anything wrong with jet holes.
4. Storm now get first connection from rad.
5. Removed the chip block from loop, don't think it does much anyway. I'll just get a quieter HS or a passive one.
6. Put pump on case floor.
7. During leak test made sure I got all air out of components, a lot of shaking going on..

Same O/C
Overclock:
CPU - Stock: 3400MHz - O/C: 3706MHz
GPU - Stock: Core 350MHz, Mem 1.0GHz - O/C: Core 410MHz, Mem 1.11GHz
DRAM Ratio: 1:1
CPU Coret: 1.3875v
DDR SDRAM: 2.7v
AGP: 1.55v


OK Temps.... We'll I'm now ecstatic about the temps I'm getting. Gone from one extreme to the other.
CPU
Probe - Idle: 26.5, Load: 30.5
MM5 - Idle: 30, Load: 37
CASE
Probe - Idle: 23.5, Load: 24.5
MM5 - Idle: 28, Load: 31
GPU
Core- Idle: 39, Load: 42
Abmient- Idle: 35, Load: 38
RAD Tube out
Probe - Idle: 24.5, Load: 28

Room Temp: 18
A few more pics
Storm
CPU Probe
Case Side
Case Close
Rad Close
Case Front - Rad

Now I'm a happy camper.

 

[morpheus6d9]

congrats on the system im looking into building a w/c system myself

 

[CrustyDemon]

Decided to add Chip block back into loop considering I've aready purchased/used it.

Loop - Rad>CPU>GPU>Chip>Res>Pump>Rad

Flow still looks really good and temperatures haven't changed.
Probe
Cpu: 26.2 idle, 31 Load
Gpu: 39 idle, 43 Load
Case: idle, 23.5
Rad Out: 25 idle, 27 Load
Room: 19

But mostly the damn chip cooler irritating noise had gone....

Have no idea why temps were so bad before, maybe just air pockets combined with bad seating cause most the problem.