![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Project: parallel HEX (completed!)
- Thread starter Mysterae
- Start date
themark said:
You got it, not quite photoshop but software of simillar ilk. Simply export two 2D images, one with transparencey and the other in full colour. Don't move the view point between image exports. Simply place both images on top of each other and edit to show what you want - very effective! Editing the image is a breeze because of the black outlines, you really can't make a mistake. Good luck on the trade show!
I'll take this oppertunity to post a little update. My water cooling stuff from Aqua Computer arrived:
As a little distraction while I wait for other parts to arrive, I'm modding the AC Cuplex XT into 3 ports, here's the thread in the Water Cooling forum about it.
Plan of action now is to benchmark the system for temps for idle and load while on air so I can compare temps when on water.
Another bit of lingo I'm not famillar with, but thanks (I think)!jakedeez said:
Erasmus354
[H]F Junkie
- Joined
- Mar 12, 2004
- Messages
- 9,450
Mysterae said:
Make sure to record ambient temps as well during the testing or it is less than useless as a comparison basis.
Duely noted Erasmus!
I've been busy taking the pc apart and test fitting the waterblocks. I have posted a preliminary review of the Aqua Computer hardware in the Water Cooling Forum that may be of interest to many.
Here's a teaser shot until the next update:
I've been busy taking the pc apart and test fitting the waterblocks. I have posted a preliminary review of the Aqua Computer hardware in the Water Cooling Forum that may be of interest to many.
Here's a teaser shot until the next update:
SpangeMonkee
[H]ard|Gawd
- Joined
- Dec 18, 2001
- Messages
- 1,435
Mysterae said:
mmmm.. :drool:
sexay!
Its time to tackle the water cooling which is the central part of my build.
The modification to the Aqua Computer Cuplex XT is complete, and I really had to test that before it went anywhere near the case, but without the Collector made, I had to improvise!
The thread about the Cuplex modification can be found here, but here are a few images to jog your memory:
The top before:
The plexi top after:
And the jet plate after modification:
The underneath of the jet before:
And after:
So all that was left was to test it:
Now, like I mentioned before, the Collector isnt made so I needed some thing else .a bottle of [H]AID anyone?
Okay, a bit g[H]etto but it works! The bottle gives an idea of what the Collector will do. But first, the acrylic manifold it works a treat at splitting the circuits. I have blocked off two outlets of the manifold because I wasnt sent enough 1/4" and 1/8 BSPP fittings to include the hard drive coolers in this test so theyll have to wait.
Here are some shots of the manifold (with flash):
Way back at the start I had two manifolds made, but I have since changed the design of the circuit to only use one. So I used the least polished one and sealed it up with clear silicone sealant, tightened it up and no leaks!
Without flash:
Now, I had to modify the pump to accept my choice of tubing (12/9mm), and that turned out well and leak free:
Ive still to mod the other pump too, so Ill post how I did that later with that pump if anyones interested. I should mention here that that strip of LEDs in the pic above are UV, kindly donated by a friend. There's some UV dye in the coolant mix to help find leaks.
Some of the other blocks in the parallel loops, the AC Aquagratix coolers for the two X1900 videocards:
The red LEDs on the Aquagratix are definitely going, Ill use them somewhere else
. Im thinking blue or UV LEDs in their place, connected to the fan header on the X1900s. This will be a good indication of the heat of the GPU; the hotter the card, the brighter the LED!
The plexi inserts that say Aqua Computer are to be replaced. Im torn between Slave and Master or Parallel and HEX .what do you guys think?
Heres the AC Twinplex Pro North Bridge cooler:
Now, lets talk about water flow. Remember that the flow in the images shown here will be different from what I will get when its all installed in the case. Firstly, because only one pump is in this circuit here and there will be two in the final installation. Also, the two hard drive coolers are not in these circuit shots, so they will take some of the flow.
In the images below you can see the flow from each return line, with the D5 pump at setting 1. Starting with the valve thats nearest you on the left of the image it switched off because its a dummy, it would have been the hard drive coolers had they been in circuit.
Going clockwise to the left, is one return of the CPU, but you cant see the flow because it is obscured. Next, looking through the bottle you can see the flow from the north bridge cooler, the AC Twinplex Pro. The next two are each of the video card coolers, AC Aquagratxs. The last flow in the image (nearest you on the right) is the other return from the CPU cooler) modded Cuplex XT. The flow from the two Cuplex outlets are both a bit less that the other waterblocks, but you would expect that since it only has one feed to it. If I close the ball valves of the other lines the flow in these CPU lines increases as you would expect. The pump at setting 1 is really silent, and doesnt create a lot of turbulence and bubbles in the bottle. There is no water splashing noise at all at this setting.
If I up the setting on the D5 pump to setting 5, things change!
Flows obviously increase, and as does the turbulence in the bottle, to the extent that the pump starts sucking the bubbles in. Not good. You can hear the air in the pump and see the bubbles flowing around each circuit. Perhaps something inside the bottle to stop the turbulence would fix that, but increasing the water level to the same as the lines feeding it in works. I cant do that with this bottle as the holes arent sealed!
I cant get a photo of the turbulence due to the long shutter time of the camera and the movement of the bubbles. The flash just bounces of the shiny surface of the bottle. Ill try again later and some of the UV in the dark.
I should have the proper Collector by the next update, and I can put the bottle of [H]AID back in the fridge!
The modification to the Aqua Computer Cuplex XT is complete, and I really had to test that before it went anywhere near the case, but without the Collector made, I had to improvise!
The thread about the Cuplex modification can be found here, but here are a few images to jog your memory:
The top before:
The plexi top after:
And the jet plate after modification:
The underneath of the jet before:
And after:
So all that was left was to test it:
Now, like I mentioned before, the Collector isnt made so I needed some thing else .a bottle of [H]AID anyone?
Okay, a bit g[H]etto but it works! The bottle gives an idea of what the Collector will do. But first, the acrylic manifold it works a treat at splitting the circuits. I have blocked off two outlets of the manifold because I wasnt sent enough 1/4" and 1/8 BSPP fittings to include the hard drive coolers in this test so theyll have to wait.
Here are some shots of the manifold (with flash):
Way back at the start I had two manifolds made, but I have since changed the design of the circuit to only use one. So I used the least polished one and sealed it up with clear silicone sealant, tightened it up and no leaks!
Without flash:
Now, I had to modify the pump to accept my choice of tubing (12/9mm), and that turned out well and leak free:
Ive still to mod the other pump too, so Ill post how I did that later with that pump if anyones interested. I should mention here that that strip of LEDs in the pic above are UV, kindly donated by a friend. There's some UV dye in the coolant mix to help find leaks.
Some of the other blocks in the parallel loops, the AC Aquagratix coolers for the two X1900 videocards:
The red LEDs on the Aquagratix are definitely going, Ill use them somewhere else
. Im thinking blue or UV LEDs in their place, connected to the fan header on the X1900s. This will be a good indication of the heat of the GPU; the hotter the card, the brighter the LED! The plexi inserts that say Aqua Computer are to be replaced. Im torn between Slave and Master or Parallel and HEX .what do you guys think?
Heres the AC Twinplex Pro North Bridge cooler:
Now, lets talk about water flow. Remember that the flow in the images shown here will be different from what I will get when its all installed in the case. Firstly, because only one pump is in this circuit here and there will be two in the final installation. Also, the two hard drive coolers are not in these circuit shots, so they will take some of the flow.
In the images below you can see the flow from each return line, with the D5 pump at setting 1. Starting with the valve thats nearest you on the left of the image it switched off because its a dummy, it would have been the hard drive coolers had they been in circuit.
Going clockwise to the left, is one return of the CPU, but you cant see the flow because it is obscured. Next, looking through the bottle you can see the flow from the north bridge cooler, the AC Twinplex Pro. The next two are each of the video card coolers, AC Aquagratxs. The last flow in the image (nearest you on the right) is the other return from the CPU cooler) modded Cuplex XT. The flow from the two Cuplex outlets are both a bit less that the other waterblocks, but you would expect that since it only has one feed to it. If I close the ball valves of the other lines the flow in these CPU lines increases as you would expect. The pump at setting 1 is really silent, and doesnt create a lot of turbulence and bubbles in the bottle. There is no water splashing noise at all at this setting.
If I up the setting on the D5 pump to setting 5, things change!
Flows obviously increase, and as does the turbulence in the bottle, to the extent that the pump starts sucking the bubbles in. Not good. You can hear the air in the pump and see the bubbles flowing around each circuit. Perhaps something inside the bottle to stop the turbulence would fix that, but increasing the water level to the same as the lines feeding it in works. I cant do that with this bottle as the holes arent sealed!
I cant get a photo of the turbulence due to the long shutter time of the camera and the movement of the bubbles. The flash just bounces of the shiny surface of the bottle. Ill try again later and some of the UV in the dark.
I should have the proper Collector by the next update, and I can put the bottle of [H]AID back in the fridge!
Emission
Supreme [H]ardness
- Joined
- Dec 6, 2005
- Messages
- 4,420
The water setup looks breautiful! I get the feeling a few elbows might help in those tight spaces
Emission
Supreme [H]ardness
- Joined
- Dec 6, 2005
- Messages
- 4,420
You only need 1 or 2 when it gets too tight, if at all. Hey, you might not need em.
rogue_jedi
Supreme [H]ardness
- Joined
- Sep 6, 2002
- Messages
- 4,821
very nicely done. i like the mods to the block. i hope they actually help
the black knight always triumphs!
the black knight always triumphs!
Emission
Supreme [H]ardness
- Joined
- Dec 6, 2005
- Messages
- 4,420
Only issue I see is that the single line feeding the 6-way splitter isn't enough to put through decent flow. I'd be a little concerned.
Thanks guys for your encouragement, comments and criticisms, it's what makes it fun to share.
After seeing the flow from each separate outlet in the bottle with the D5 pump on setting 1, I was a little disappointed. However, at setting 4 and 5 there is plenty of flow and that's only with one pump.
Flowing water isn't something you normally see in pc water cooling set ups, so it's hard to gauge what's acceptable flow or not. I guess I won't know until some heat load is put on each block and the temps monitored.
The water going into each port in the collector has got to be under the water level. That's something I have learned from setting up the test circuit with the bottle. Not only does water falling from a height create turbulence and bubbles, when the pump is switched off the water in each line to the bottle falls back to the exact water level in the bottle. It's the same thing as using a long tube filled with water to find the level in room with large walls, before laser levels of course. It's no big deal really, but it's wierd to watch the water go 'backwards' when you turn off the pump!
Unfortunately I have to take a week off from the mod due to life getting in the way. Here's what's still to be done:
- wait for pumbing order for the out of stock items
- have the Collector made
- finalise the case construction, I had to revise the design due to frame strength, aesthetics and replacement of certain components.
- submit the case designs for construction
- mod the power supply load LED
- mod the gpu coolers LEDs
- etch something on the gpu acrylic inserts
- mod the other pump to accept the tube diameter of choice
- mod both pumps control pot to a dual gang potentiometer
- sleeve exposed additional wiring
- mod the Cooler Master crossflow fans into clear acrylic tubing (this will look great)
- wire up the usual buttons and fan control
- construct the case
I'm tempted to mod the AC Aquagratix gpu coolers that may be a bit more in keeping with the rest of the case, more on that later.
I think I should be finished by mid May, 2006 of course!
After seeing the flow from each separate outlet in the bottle with the D5 pump on setting 1, I was a little disappointed. However, at setting 4 and 5 there is plenty of flow and that's only with one pump.
Flowing water isn't something you normally see in pc water cooling set ups, so it's hard to gauge what's acceptable flow or not. I guess I won't know until some heat load is put on each block and the temps monitored.
The water going into each port in the collector has got to be under the water level. That's something I have learned from setting up the test circuit with the bottle. Not only does water falling from a height create turbulence and bubbles, when the pump is switched off the water in each line to the bottle falls back to the exact water level in the bottle. It's the same thing as using a long tube filled with water to find the level in room with large walls, before laser levels of course. It's no big deal really, but it's wierd to watch the water go 'backwards' when you turn off the pump!
Unfortunately I have to take a week off from the mod due to life getting in the way. Here's what's still to be done:
- wait for pumbing order for the out of stock items
- have the Collector made
- finalise the case construction, I had to revise the design due to frame strength, aesthetics and replacement of certain components.
- submit the case designs for construction
- mod the power supply load LED
- mod the gpu coolers LEDs
- etch something on the gpu acrylic inserts
- mod the other pump to accept the tube diameter of choice
- mod both pumps control pot to a dual gang potentiometer
- sleeve exposed additional wiring
- mod the Cooler Master crossflow fans into clear acrylic tubing (this will look great
)- wire up the usual buttons and fan control
- construct the case
I'm tempted to mod the AC Aquagratix gpu coolers that may be a bit more in keeping with the rest of the case, more on that later.
I think I should be finished by mid May, 2006 of course!
Wow, keep up the good work. I liked the idea of the water falling from the tubes into the collector though. That would have been incredible-looking. Have you considered trying to add a filter or something of the sort to get rid of the bubbles? If not, maybe you could make the res so deep or so big that the bubbles would be gone before they reached the pump's feed. The water going "backwards" shouldn't be a big deal, especially if the ends of the circuit drain at a higher altitude than the components of the build.
I agree, the sight of water tumbling down is impressive. We are discussing the elimination of the bubbles in the cuplex thread in the water cooling forum at the moment. I was thinking pin-pong ball or plexi disc, I feel some more experimentation!
You would be surprised at how deep the bubbles reach. The lowest setting of the pump when bubbles aren't created is between 2 and 3, and it mixes the water up real good. At that level tiny bubbles get created and get sucked in to the circuits.
You are right about the backflow not being a big deal, all water blocks are lower than the reservoirs water level so it doesn't impact it. The radiator is highest, but the pressure from the water mass keeps that filled.
You would be surprised at how deep the bubbles reach. The lowest setting of the pump when bubbles aren't created is between 2 and 3, and it mixes the water up real good. At that level tiny bubbles get created and get sucked in to the circuits.
You are right about the backflow not being a big deal, all water blocks are lower than the reservoirs water level so it doesn't impact it. The radiator is highest, but the pressure from the water mass keeps that filled.
Gongo said:
If you look closely, he has a valve on each loop, so he can adjust the flow as necessary. But I really doubt it will be a problem, the pump is plenty strong enough to move fluid through all 6 loops, and as long as there is any flow at all, it will cool far better than Air will.
==>Lazn
Thanks guys!
Nothing of the case has been made from plexi yet, still a bit of working out to do.
I have changed a few things from the original mock-up - it depends on if I can get away with one radiator or need another one. If I only need the one rad, it will go up top as per the original and the hard drives will be relocated to the rear of the case. If I do need two rads, then these will go at the rear of the case and the hard drives will stay on the side panel.
So I need to see how well the parallel water cooling stands up to stress before I can finalise the construction of the case. Here are my current air temperatures:
Air Cooling Temperatures
*AMD X2 4400 Overclocked to 2.64Ghz on stock AMD HSF.
**When using CCC Overdrive, only the X1900XTX is overclocked, and at 690Mhz Core, 800Mhz Mem. Fan control for both cards is on auto.
***Both cards are overclocked using ATITool. X1900CF @ 670Mhz Core, 801Mhz Mem and X1900XTX @ 725Mhz Core, 891Mhz Mem. Fan speeds were set to an ear bleeding 100%.
Obviously I want as low temperatures as possible, but my target load temps are below 45C for the CPU, and below 40C for both gpu's while retaining overclocks. If I can get even better overclocks with water than air, that's a plus!
Nothing of the case has been made from plexi yet, still a bit of working out to do.
I have changed a few things from the original mock-up - it depends on if I can get away with one radiator or need another one. If I only need the one rad, it will go up top as per the original and the hard drives will be relocated to the rear of the case. If I do need two rads, then these will go at the rear of the case and the hard drives will stay on the side panel.
So I need to see how well the parallel water cooling stands up to stress before I can finalise the construction of the case. Here are my current air temperatures:
Air Cooling Temperatures
*AMD X2 4400 Overclocked to 2.64Ghz on stock AMD HSF.
**When using CCC Overdrive, only the X1900XTX is overclocked, and at 690Mhz Core, 800Mhz Mem. Fan control for both cards is on auto.
***Both cards are overclocked using ATITool. X1900CF @ 670Mhz Core, 801Mhz Mem and X1900XTX @ 725Mhz Core, 891Mhz Mem. Fan speeds were set to an ear bleeding 100%.
Obviously I want as low temperatures as possible, but my target load temps are below 45C for the CPU, and below 40C for both gpu's while retaining overclocks. If I can get even better overclocks with water than air, that's a plus!
Everfalling
Gawd
- Joined
- Jul 12, 2005
- Messages
- 676
i'm just wondering, where did you get that thermo camera?
SpangeMonkee
[H]ard|Gawd
- Joined
- Dec 18, 2001
- Messages
- 1,435
bail_w said:
with the processor being overclocked and running on a stock cooler, I don't think it is too bad.
board2death986
[H]ard|Gawd
- Joined
- Aug 13, 2005
- Messages
- 1,404
When is the collector going to be finished?? This project is [H]ard as hell, love the water cooling setup, and the fact you took the time to prefab the entire case to spec before going with the final product. This is encouraging me to attempt my own watercooling setup, but I will need to do serious research and test on a junk system beforehand. Do you know if the non-conductive fluid dissapates heat as well as coolant?
I agree with SpangeMonkee, I thought the temp of the CPU (58C) was pretty good considering it's the stock cooler, overclocked by 440Mhz, and running two instances of Prime95. That's probably the worst case scenario.
I borrowed the thermal camera from work. I'm sure you can hire them, but from what I learned from using it you need an appreciation of emissivity. For instance, take two identical materials at the same temp, but paint one mat black and the other normal. They will show up with different temperatures on the camera. The way around this and the method I will be using once the case is built, is to place black electrical tape on the components you want to measure. Enter the emissivity for black tape into the camera and you will have accurate temps.
board2death986, the Collector should be finished by next week, it all depends on materials coming in and my machineshop mate to be free. As for non-conductive fluid, it's best if you ask in the watercooling forum as I've never used it and not qualified to say. I'm only using summer coolant and distilled water as all connections will be tighened up and I don't foresee any leaks. The test with the bottle and blocks has proven to be leak free so far. Yay!
I borrowed the thermal camera from work. I'm sure you can hire them, but from what I learned from using it you need an appreciation of emissivity. For instance, take two identical materials at the same temp, but paint one mat black and the other normal. They will show up with different temperatures on the camera. The way around this and the method I will be using once the case is built, is to place black electrical tape on the components you want to measure. Enter the emissivity for black tape into the camera and you will have accurate temps.
board2death986, the Collector should be finished by next week, it all depends on materials coming in and my machineshop mate to be free. As for non-conductive fluid, it's best if you ask in the watercooling forum as I've never used it and not qualified to say. I'm only using summer coolant and distilled water as all connections will be tighened up and I don't foresee any leaks. The test with the bottle and blocks has proven to be leak free so far. Yay!
board2death986
[H]ard|Gawd
- Joined
- Aug 13, 2005
- Messages
- 1,404
is the collector done yet??? anxious to see how this mod finishes up.
Unknown-One
[H]F Junkie
- Joined
- Mar 5, 2005
- Messages
- 8,909
All you need to do is put a cone in the middle of the collector to fix both the water splashing noise and the bubbles...the hard part will be aiming all the strams to fall on the top of the cone so that all the water will run down it evenly.
Apologies to anyone thats been waiting on an update, real life got in the way again as it invariably does. Sadly I dont have the Collector made yet, I had to finalise a few things and fundamentally test my parallel water cooling with hot full powered hardware. This was to make sure the cooling would do what I hoped it would do, and also to see if I needed another 2.120 radiator. More on all this in a bit!
dchrsf - I had thought about upgrading the motherboard to a DFI CFX-3200 now that its available, but I have the RDX200 running very nicely. Apart from craptacular USB performance, it has bedded in to be a really good board. The problem I was having with stuttering during high performance 3D was due to a rogue network port that when disconnected put a huge demand on the CPU causing the performance loss. Disabled in the BIOS, its no longer a problem. Having said all the above, Im really curious the jump in performance a PCIe 2 x16 motherboard would give me. Thats an idea for anyone wanting to write a review compare the DFI RDX200 against the DFI CFX3200 with Crossfire X1900 Id love to do it, but the price of the CFX200 is terrible at the moment with (my perceived) little gains, so not at this time as they say.
Unknown-One your cone in the Collector is an excellent idea! Incidentally, this has happened by accident. After seeing the flow with one pump and being a little disappointed, adding in the second pump of course increased the flow, but it also means the streams of water will all meet at one single point in the Collector the DD Fillport that is fitted in the centre of the Collector. With the correct water level, the water streams will join at the fillport and kill any turbulence that may occur. To be able to use my Ghetto Sprite Collector, I used some polystyrene packing stuff at the water surface to kill the turbulence and it works a treat. Of course this will not be fitted to the final build!
On to the update who said water and electronics doesnt mix:
Everything except the hard drives is now water cooled! To think that all that has to go in to a case .its going to weigh a ton!
Some more pictures to focus in on certain aspects.
The main parts of my idea of water cooling the radiator, Ghetto-Sprite-Collector, distributor etc. Not all the connections are tightened up as the 12/9mm tubing (for this section) fits quite well and is only temporary for testing.
Here are the two modified D5 pumps providing the flow. These are placed before and after the radiator.
The Ghetto-Sprite-Collector I must be mad to run live with this thing! But hey, its a means to an end and works well. You can see the polystyrene packing I stuffed in there to kill the turbulence below the valve line, and it does the job. It makes the water bubble at the edges, but I can assure you the water isnt boiling! The droplets you see above the valve line is not due to splashing of any kind, its due to the condensation that occurs, people with acrylic reservoirs will know of this.
The CPU and North Bridge water blocks (modified AC Cuplex XT and AC Twinplex respectively). The modified Cuplex is working well, Ill show you some temperatures in a bit. The mounting for the Twinplex could be better; I dont think the spring clips I have used from the stock HSF are giving it enough downwards pressure.
X1900 Crossfire water cooled AC style:
The temps from these coolers arent phenomenal, probably due to them cooling the GPU, memory and power ICs at the same time. But 70C at full overclocked load, very little surrounding airflow, quieter than a library all at an ambient temp of 27C is okay for me!
Speaking of temperatures, here we go:
As a reminder here are the temps on air, very noisy air to be precise:
Now the water cooled temps, with ONLY ONE D5 pump at the 5 setting:
And now with TWO PUMPS both at the 5 setting:
Before analysing the results, there are some health warnings that apply here that may or may not apply to the final build in the case. Firstly, the ambient temps are different, much higher in the 3rd run of tests. Its somewhat due to the warmer weather weve been having here, not that Im complaining about that, and its good that it has happened as it reflects the real world temps later in the year (heres hoping). Its also due to the amount of testing I performed the longer the system ran, the more heat it dumped into my room. I will re-run the tests with cooler air coming into the room for better temps, but all tests here were done with the window closed.
Then theres the amount of tubing I have and the amount of water in the Ghetto-Sprite-Collector. The last time I emptied it, the Sprite bottle had over a litre of coolant in it. Thats way too much, and gets lukewarm during heated testing. In the final build the collector will have less than a can of coke and there will be a lot less tubing involved.
Lastly, the airflow is a little different, due to no moving air from the CPU, NB and GPUs HSFs (heat sink fans). It would be nigh impossible to recreate the kind of airflow they produce while water cooling. The crossflow fans were kept at the same minimum low speed for all testing.
While waiting for the Collector to be built for final testing (hurry up goddamit!!), I have some time to contemplate the need for another radiator. I feel that is a thread in its own right, so I have created one in the Water Cooling section, feel free to drop by and give any advice you see fit.
dchrsf - I had thought about upgrading the motherboard to a DFI CFX-3200 now that its available, but I have the RDX200 running very nicely. Apart from craptacular USB performance, it has bedded in to be a really good board. The problem I was having with stuttering during high performance 3D was due to a rogue network port that when disconnected put a huge demand on the CPU causing the performance loss. Disabled in the BIOS, its no longer a problem. Having said all the above, Im really curious the jump in performance a PCIe 2 x16 motherboard would give me. Thats an idea for anyone wanting to write a review compare the DFI RDX200 against the DFI CFX3200 with Crossfire X1900 Id love to do it, but the price of the CFX200 is terrible at the moment with (my perceived) little gains, so not at this time as they say.
Unknown-One your cone in the Collector is an excellent idea! Incidentally, this has happened by accident. After seeing the flow with one pump and being a little disappointed, adding in the second pump of course increased the flow, but it also means the streams of water will all meet at one single point in the Collector the DD Fillport that is fitted in the centre of the Collector. With the correct water level, the water streams will join at the fillport and kill any turbulence that may occur. To be able to use my Ghetto Sprite Collector, I used some polystyrene packing stuff at the water surface to kill the turbulence and it works a treat. Of course this will not be fitted to the final build!
On to the update who said water and electronics doesnt mix:
Everything except the hard drives is now water cooled! To think that all that has to go in to a case .its going to weigh a ton!
Some more pictures to focus in on certain aspects.
The main parts of my idea of water cooling the radiator, Ghetto-Sprite-Collector, distributor etc. Not all the connections are tightened up as the 12/9mm tubing (for this section) fits quite well and is only temporary for testing.
Here are the two modified D5 pumps providing the flow. These are placed before and after the radiator.
The Ghetto-Sprite-Collector I must be mad to run live with this thing! But hey, its a means to an end and works well. You can see the polystyrene packing I stuffed in there to kill the turbulence below the valve line, and it does the job. It makes the water bubble at the edges, but I can assure you the water isnt boiling! The droplets you see above the valve line is not due to splashing of any kind, its due to the condensation that occurs, people with acrylic reservoirs will know of this.
The CPU and North Bridge water blocks (modified AC Cuplex XT and AC Twinplex respectively). The modified Cuplex is working well, Ill show you some temperatures in a bit. The mounting for the Twinplex could be better; I dont think the spring clips I have used from the stock HSF are giving it enough downwards pressure.
X1900 Crossfire water cooled AC style:
The temps from these coolers arent phenomenal, probably due to them cooling the GPU, memory and power ICs at the same time. But 70C at full overclocked load, very little surrounding airflow, quieter than a library all at an ambient temp of 27C is okay for me!
Speaking of temperatures, here we go:
As a reminder here are the temps on air, very noisy air to be precise:
Now the water cooled temps, with ONLY ONE D5 pump at the 5 setting:
And now with TWO PUMPS both at the 5 setting:
Before analysing the results, there are some health warnings that apply here that may or may not apply to the final build in the case. Firstly, the ambient temps are different, much higher in the 3rd run of tests. Its somewhat due to the warmer weather weve been having here, not that Im complaining about that, and its good that it has happened as it reflects the real world temps later in the year (heres hoping). Its also due to the amount of testing I performed the longer the system ran, the more heat it dumped into my room. I will re-run the tests with cooler air coming into the room for better temps, but all tests here were done with the window closed.
Then theres the amount of tubing I have and the amount of water in the Ghetto-Sprite-Collector. The last time I emptied it, the Sprite bottle had over a litre of coolant in it. Thats way too much, and gets lukewarm during heated testing. In the final build the collector will have less than a can of coke and there will be a lot less tubing involved.
Lastly, the airflow is a little different, due to no moving air from the CPU, NB and GPUs HSFs (heat sink fans). It would be nigh impossible to recreate the kind of airflow they produce while water cooling. The crossflow fans were kept at the same minimum low speed for all testing.
While waiting for the Collector to be built for final testing (hurry up goddamit!!), I have some time to contemplate the need for another radiator. I feel that is a thread in its own right, so I have created one in the Water Cooling section, feel free to drop by and give any advice you see fit.