Water Cooling Circuits Analysis (A liquid cooling work log)

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Thu Sep 02, 2004 9:13 pm

Scenario 3 is up and running; finished both passes of Test One today and it's looking quite good. Here're the customary images.

Image
Here's the new line running out of the reservoir.
A 3/8" y-splitter divides flow up to the two pumps.

Image

Image
Here are the two pumps, no longer joined directly as in Scenario 2.


Image

Image

Image
A full loop view.


Image
The new line from the second pump, going into the CPU block.


Image
The new line from the first pump, going into the northbridge block before flowing into the GPU block.


Image
The new line heading into the reservoir, where the two pump/block paths recombine.


Image


I should have the remaining tests done, hopefully, before Monday, so that I can post the final results of the circuit analysis.

I have decided that I would like to take this little circuit analysis a step further, and later on continue this thread with an analysis of how my upcoming upgrades affect the performance of this system. The first upgrade will be replacing the single D-Tek Pro-120 heatercore with a pair of Black Ice Pro heatercores, one right after the pump, before the block(s), in each branch of the parallel loops. We will get to see how dual 1" thick 120mm heatercores compares in performance to a single 2" thick 'core.

After that, I will be upgrading from the MCW-6000A to one of Cathar's upcoming, "Storm," (only a codename right now) G4 blocks. I will be able to also analyze the performance improvement by going from the MCW-6000A to his new block.

Actually, thrown in somewhere in there, will be a comparison of the Mark I CSP-750 pumps to the Mark II models, even though it's not really important, since Mark II effectively replaces, rather than sells alongside, the Mark I model.

I guess we can say that this thread will turn into my own water cooling worklog, of sorts.

-Ed

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Fri Sep 03, 2004 1:54 pm

why not get another heater core and set one pump per core and use one for the cpu only and one for the northbridge and gpu?

and put like 2 tlines in stead of a shared res. I wonder if it owuld work much better.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sun Sep 05, 2004 9:09 am

Okay, everyone. One last circuit-related update, and then on to other things. Here are the results from all three circuit arrangements:

Image

Image

Image

As we can see, it appears that the dual pumps in parallel loops performed the best, but only by a miniscule margin, which can even be to be within the margin of error for my slack testing procedure; however, there's two things to note:

Firstly, even if it is within the realm of margin of error, the Delta-T is still always in the no change to temp. reduction range with the exception of the CPU in Test 2 and Northbridge in Test 3, but that chip has been perfectly comfortable in all testing, anyway. All my tests for Scenario 3 were performed so that I waited until they seemed to stabilize, and then let it go for an extra six hours to be safe. This resulted in tests taking anywhere from around seven hours to up to nine hours.

Secondly, and more importantly, this sets me up quite well for the second phase of circuit testing. My second phase? Simple...

I already have in my hands two brand new Black Ice Pro radiators. I did a bunch of reading on heatercore size and shape as well as the important of air flow resistance and pressure drop with low pressure fans such as those we use in modern PCs (as opposed to the much higher pressure, higher velocity blower types that most automotive heatercores are designed to work with) and decided that I would like to see what happens when I transition from a 2" thick single heatercore with wavey, higher turbulence veins to a pair of ~1" thick, smooth vein radiators. Because order of components in the loop with exception of what comes before the pump is mostly irrelevant (at least with the accuracy of my testing equipment), I will be comparing the following two arrangements:


1) Current setup:

Reservoir -> Y-Splitter -> A/B:

A: Pump1 -> Chipset Block -> GPU Block -> Y-Combiner

B: Pump2 -> CPU Block -> Y-Combiner

Y-Combiner -> Pro-120 Heatercore -> Reservoir


versus:


2) New setup:

Reservoir -> Y-Splitter -> A/B:

A: Pump1 -> Chipset Block -> GPU Block -> Black Ice Pro2 -> Y-Combiner

B: Pump2 -> Black Ice Pro1 -> CPU Block -> Y-Combiner

Y-Combiner -> Reservoir


The entire system will be drained and all components will be removed in order to perform case modification in preparation for switching to the new setup. Case modifications include drilling out the proper holes for direct mounting of the radiators to the case, with merely four holes in the front case wall, but major modification to the case rear wall, where I will widen the exhaust hole downwards, create new mountholes for the radiator to hardmount, and then cut out the necessary material above the new air pathway to allow the radiator's fittings to poke in through the rear wall. This will allow me to have the second radiator mounted externally without running any tubing out of the case itself.

I'll be utilizing CoolWorks' Mini-CoolShrouds for the two Black Ice Pros. Black Ice Pro1 will be directly mounted to the front intake, with the shroud behind it and then the fan behind that, pulling air through; it will be the same fan as currently used for the Pro 120 and its fittings will be in the same orientation as the Pro 120's. Black Ice Pro2 will be mounted externally out back directly to the new exhaust port, with the Mini-CoolShroud further behind it and then the current Antec 120mm blue LED fan mounted behind the shroud, once again pulling air through. The fittings will actually go through a new opening in the rear case wall right into the case, entering under the PSU.

The side panel fan will be flipped again so that it's pulling in cold air once more, to help alleviate some of the temperature difference between two radiators. There is also a distinct possibility that I will be swapping out the Fortron FSP300-60PN for a PSU of the SeaSonic variety.

Let me know what you guys think. Now that I finally have this circuit comparison wrapped up, I can slow down and take care of some reviews that have gone neglected due to this test. I can take my time figuring out exactly how to go about comparing the two radiator arrangements, so feel free, now, to suggest some more complex modifications that will probably be doable, now that I'll be removing everything from the case, and now that we know what radiators will be replacing the Pro 120. Just keep in mind that portability remains of prime importance.

-Ed

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Sun Sep 05, 2004 9:35 am

very interested to see how this turns out.

how about a cost tally?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sun Sep 05, 2004 9:37 am

Cost tally? I think I'm gonna' cry. :oops:

1911user
Posts: 109
Joined: Sat Mar 13, 2004 12:08 pm
Location: Oklahoma

Post by 1911user » Sun Sep 05, 2004 8:05 pm

Nice work Ed! You'll be a vetern water cooler soon.

If you get one of Cathar's waterblocks, you'll be set to watercool another computer!
cpu block (check)
pump (check)
rad (check)
tubing (probably)
fan (a given)
shroud (K-mart :))

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Mon Sep 06, 2004 12:38 pm

:lol:

I guess considering that my costs amounted to over two whole systems' worth of parts, then maybe they're not as bad as I suppose. A lot of the additional cost also arose from major mistakes I made...

If I get some time later, I'll go ahead and whip up a cost guesstimate.

And thank you, 1911.

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Mon Sep 06, 2004 4:58 pm

ed at least you have that kind of money to burn.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Mon Sep 06, 2004 5:07 pm

Young but working, yet living at home. We decided it best, because the money saved in bills and rent by me not living outside allows us to more rapidly build capital for the restaurant. Part of my earnings go to me, and part of it goes to my parents. Certain things are up in the air right now, like who may or may not move to Florida and when, because the new restaurant is down there, but dad has one here in NY and he's not about to just let it go.

Rather off topic, so I'm not really going to discuss any more of it here. Let's just say I don't have siginificant monetary responsibilities like my own bills, on top of other things, so it makes things quite a bit easier. No car of my own, no car insurance, no rent, no this bill, no that bill (except for my school loans; I pay for those myself). I know what it means to be my own entity; now is not the time to take up something like that.

-Ed

Fastman
Friend of SPCR
Posts: 47
Joined: Mon Oct 28, 2002 3:43 pm
Location: Vancouver, BC, CANADA

1 pump / 2 cooling loops in parallel ??

Post by Fastman » Mon Sep 06, 2004 7:22 pm

I was wondering how you think the system would work with only 1 pump and and 2 parallel cooling loops?

This would allow the one pump to pump against less resistance and provide more flow than all cooling blocks in series.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Mon Sep 06, 2004 7:29 pm

I honestly don't entirely know without actually trying it out; however, there is one thing that I think is important if this were to be tried...

The pump's in and out lines should be broader than the two parallel branches. In other words, for example, the reservoir should have a 1/2"ID outlet to the pump, and the line from the pump to the Y-splitter should also be 1/2"ID, but then the two branches out of the Y should be 1/2"OD.

The truth is, however, as I said, this is impossible for me to say without actually trying it, as I do not know the necessary arithmetic to, "simulate," the effects.

-Ed

1911user
Posts: 109
Joined: Sat Mar 13, 2004 12:08 pm
Location: Oklahoma

Re: 1 pump / 2 cooling loops in parallel ??

Post by 1911user » Mon Sep 06, 2004 8:00 pm

Fastman wrote:I was wondering how you think the system would work with only 1 pump and and 2 parallel cooling loops?

This would allow the one pump to pump against less resistance and provide more flow than all cooling blocks in series.
Each block would see less flow in parallel loops. Having everything in series is almost always the best arrangement for normal pumps and blocks. The exception is when a component is extremely restrictive and not cooling something really hot (hdd, NB, etc.), then a parallel loop with a restrictive branch or 2 might make sense. Those generally occur when someone is watercooling almost everything in the computer.

Gooserider
Posts: 587
Joined: Fri Aug 01, 2003 10:45 pm
Location: North Billerica, MA, USA
Contact:

Post by Gooserider » Tue Sep 07, 2004 7:14 pm

A couple of comments:
1. Re the pump only tests - You probably read close to ambient on those tests becase w/o the mobo heat you weren't maxxing out the rad, so it was able to get rid of all the heat that the pumps were putting out. (FWIW, my guess is that the pumps were basically adding their rated wattage to the coolant, but the rad was able to dissipate it enough that you couldn't pick it up w/ your equipment.

2. It is less than optimal to have both rads effectively in series airflow, as the 2nd rad will be fed with the warm air from the 1st one. However, I suspect your airflow would be better in the setup you describe if you just blocked off the side panel fan vent and otherwise sealed up the case. (It would be worth a test for sure)

With a sealed case, you are putting the two fans in effective series, which increases static pressure. Since the biggest problem w/ a rad is overcoming static pressure, maximizing static pressure will give the biggest boost to airflow.

With the side fan blowing in, you increase the backpressure on the front rad fan and greatly reduce it's airflow. This will aid the rear fan some, but not enough to be a net gain.

3. General theory suggests that having seperate loops entirely is better than having partially combined loops. I would attempt to make the two rads each part of their own loop. However your proposed approach isn't all that bad, except that I would rather see the pumps in series. (Pumps are best in series, rads are best in parallel)

Gooserider

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 07, 2004 7:54 pm

My proposed arrangement has the two pumps and rads in parallel, circuit-wise, but yes, the rads would be serial, airflow-wise...

I think I can try the sealed-off fan idea. Just need to figure out how to seal it out without hardcore fuglification, in the case that it works out better in terms of temps.

There is still a possibility of making the pumps in serial and leaving only the blocks & rads in parallel branches within the loop. The only problem is that if I were to try this, I would like to have the line to the pumps from the reservoir a 1/2"ID line, but then when the line out of the pumps splits up, I would like a Y-splitter that has, effectively, a 1/2" inlet, but the outlets are 3/8". I looked about, but could not find such a y-splitter anywhere, so my only choice, then, would be 3/8" everywhere, as it is now. At least with the pumps in parallel as well, I could run 1/2"ID all the way to the pumps' inlets, and then have everything from the pumps' outlets on be 3/8"ID. See what I mean?

-Ed

Gooserider
Posts: 587
Joined: Fri Aug 01, 2003 10:45 pm
Location: North Billerica, MA, USA
Contact:

Post by Gooserider » Wed Sep 08, 2004 1:49 pm

Check w/ McMasters on the splitter, I think they may have them. Otherwise I would run 1/2" from the res to the inlet of the 1st pump, possibly between the pumps, but 3/8" on the outlet from the 2nd pump on.

For sealing off the fan, I'd just use a peice of cardboard and some tape first, then if it worked try something like making a fan opening shaped plug out of plexi to get a 'window' sort of effect. Maybe get something decorative engraved on it to make it look like a design element instead of a patch?

Gooserider

Becks
Posts: 76
Joined: Mon Mar 01, 2004 11:19 pm

Post by Becks » Mon Sep 13, 2004 8:56 pm

"My proposed arrangement has the two pumps and rads in parallel, circuit-wise, but yes, the rads would be serial, airflow-wise... "

I'd watch what you do when putting rads in series with respect to the airflow.. maybe I'm not exactly sure what you're testing. I'd say its probably never a good idea unless you have enough airflow (doesn't happen with silence pcs).

I have 2 140x140 heatercores each with their own fan, they run in parallel in the water loop (whitewater block has 2 outs, each goes to a seperate rad, after the rads is a Y back into the pump and then to the cpu block)

Anyways, running the fans ~5v... the difference between the air exiting the rad and the water temp is well within the accuracy of my temp probe, 1/2 the time it says the rad is cooler than the air leaving it.

Basically that means the air exiting the rad is holding as much heat as possible, if my rads were in series airflow-wise teh 2nd rad would do absolutely nothing.

Maybe I just don't see what the point of knowing flat fins would be better... perhaps 2 blackice rads work better with flat fins vs 1 heatercore with wavy fins, but 2 heatercores would work better than 2 blackice rads. Might be interesting results but I just don't see the application.

BTW If you had enough airflow the air exiting hte rads is still below the watertemp, however this is a silentpc place... and I'd think it'd make sense to stick with low airflow because thats what would actually be used.



If you want a test that would be useful, to alot of people imo, would be running a watercooling loop, then adding like 3 90degree fittings and running it again. 90degree fittings are known to hurt flowrates, but how much exactly would it change load temps would be interesting, and useful.

I'd say one of the worst parts about watercooling is running the tubing in the case, not exactly hard, just a pain to do. When placing a rad somewhere you have to think where the tubing is going to run, trying to get it all in teh case wihtout kinks, and when your done you have a case that looks like a mess. Big swooping bends suck :P hehe, plus overall the tubing is always in the way when trying to reach other things.

I think it'd be neat to see how much using 90degree or 45degree fittings affect performance... 1-2C higher load temps would be handy information. I'd reconsider my setup (i doubt it would be that much).

Actually I got some MarkIIs on order, and when I get them I plan on redoin my case and I'm going to try using some 90s because it would really make everything eaiser. maybe just 45s over 90s but something do deal with all these big swooping bends I have to have currently.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Mon Sep 13, 2004 9:16 pm

This machine is not geared towards super low noise; as I've said a few times now, I'm in fact running both 120mm fans at their full 12volts; the machine is for heavily overclocked, portable/LAN gaming with headphones. Gaming with closed headphones, the noise limit is much higher, so I'm not running the fans low speed.

The concern over the second radiator's temperature is also why I originally planned to have the side panel fan flilled around as an intake, pulling in cold air. The exhaust pressure from the PSU fan should help relieve the possible backpressure that would have occurred should I have flipped the side fan to intake without anything but the rear 120mm fan pulling air outwards.

Actually, I am now contemplating buying a second pair of CSP-750 MarkIIs, for a total of four, believe or not. This way I can have two parallel branches, with (a) block(s), a radiator and also two pumps in serial in each branch; the NV-68 GPU block, and much more in particular, the Storm G4 that I plan to upgrade to, will benefit from having two serial pumps pushing right into them. Moreover, I have devised a way to take advantage of the small size of the CSP-75- pumps so that I can literally run the tubing straight out from the reservoir to the pumps, somehow suspended directly over/across from the blocks (try to picture this with me; I'm still figuring out exactly how to hold the blocks in these positions...), and then running the tubing directly from the pump outlets to the block inlets (all nearly perfectly straight tube runs with no bending; let's say, no more than perhaps 10 degree difference in angle between the two ends of the tube runs).

For the section from the reservoir to the pumps, I have two possible arrangements. The first arrangement would be to get a new reservoir, one of the Typhoon units, with dual inlets and dual outlets. I could then run dual 1/2"ID lines to the first pump in each branch, then another 1/2" line from the first pump to the second pump, and then 3/8" out of the second pump's outlet to the blocks, and from there on continue in 3/8" fashion right back to the reservoir, with both reservoir inlets being 3/8". The other option is to keep my current reservoir, but run a 1/2"ID line out to a special y-splitter. The problem is that I am having a tremendous amount of difficulty trying to find the y-splitter. Perhaps somebody here in the forums may be able to help me find it? I am looking for a nylon barbed fitting y-splitter, with a 1/2"OD inlet and two 3/8"OD outlets. Then with this, I can run 3.8"ID lines to the pumps. I could consider doing this with a 1/2"in/out y-splitter, since the pumps can be fitted with different size fittings for in/out as well, but I feel as though 1/2"in and 3/8" for the two outs would be beneficial. Maybe I'm wrong there, though.

I'm still trying to work out how I will suspend the pumps; I have an idea for how to orient the pumps so that they run a straight line out from the reservoir, into the pump, and then out of the pump straight into the blocks, but that's with one pump per parallel branch. If I decide to do dual pumps in serial in each parallel branch, then I need to figure out how to fit and suspend four pumps in mid-air between the reservoir and blocks, while still maintaining nothing but straight tube runs in that portion of the loop (i.e. taking advantage of the 90-degree difference in angle between inlet and outlet on the pumps).

All of these pump suspension arrangements with direct-from-res and direct-to-block tube runs seem possible only with these little CSP-750s, and this is part of the reason why I want to try this. I want to do something that's not only different, but effective. Something that I have yet to see anyone else do, if at least only because other pumps are just way too big to even consider something like this.

-Ed

Straker
Posts: 657
Joined: Fri Jul 23, 2004 11:10 pm
Location: AB, Canada
Contact:

Post by Straker » Tue Sep 14, 2004 1:04 am

not sure exactly what you're saying... so the pumps will basically be used as (or take the place of, rather) small sections of tubing and you'll try to connect them directly to blocks/the res etc?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 14, 2004 5:09 am

If you look at the way the system is arranged now, it is similar to most other setups; there are long, curving tube runs from the reservoir all the way down to the pump, sitting at the floor of the case, then another long, curving tube run back up to the block inlet. The tube goes through a 90-degree bend on the way down to the pump, and then another 90-degree bend on the way up and into the block. I plan to utilize the natural 90-degree difference in angle between the pumps' inlet and outlet, plus their small size and light weight, and suspend them across from the reservoir, directly over the block, with the pump inlet pointed right to the reservoir, and the pump outlet pointed right to the block inlet, thus allowing me to use the shortest, straightest possible tube runs to connect the reservoir to the pumps and the pumps to the first blocks in their branches.

Make sense yet? If not, I may end up having to draw some sort of diagrams, but this is very difficult for me, not normally working in three dimensions...

-Ed

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Tue Sep 14, 2004 12:32 pm

are you thinking like attaching a pump (or two) directly to the res?

that is an mcw6000 you're running right?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 14, 2004 12:57 pm

No, not directly to the reservoir, but with maybe 2-4" of tubing, almost perfectly straight, from the res to the pump inlets, and then about 2-4", almost perfectly straight, right into the block inlets from the pump outlets.

It's an MCW-6000A right now, but the moment Cathar begins to accept orders for the Storm, I'm ponying up some ca$h.

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Tue Sep 14, 2004 2:33 pm

i think 4 pumps would be a bit of overkill.

what if you put one pump after the res and the other after the cpu wb.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 14, 2004 2:50 pm

There's two parallel branches here; if I used two pumps, one pump in front of the CPU block and one after, then what happens to the GPU block and northbridge block that are on the second branch (one branch for pump, radiator and CPU block, and second branch for pump, radiator, GPU block and northbridge block)?

Having one pump after the CPU block serves no benefit over both pumps being before the block, and as a matter of fact, would be worse. The idea is to maximize the amount of pressure into the Storm G4, which is designed to take advantage of higher pressure than the MCW-6000A, which, while very performing extremely well (if not pretty much better than anything else) at low pressure and flow rates, flatlines at higher flow/pressure, while the Storm continues to reap major benefits. The new arrangement is to take advantage of the Storm's properties.

There is also a possibility of running dual pumps on the CPU block branch, but only one pump on the branch for the other two blocks, since the serial blocks would increase pressure more than flow in their branch, and the Storm is more restrictive than the NV-68 or Maze4 Northbridge blocks, but symmetrical arrangement would seem better to me, and I don't mind the cost of four pumps over three, anyway. If anything, the more complex dual-block branch would benefit from the second pump in serial to help equalize pressure and flow going in and coming out of the y-splitter/combiner into and out of the reservoir.

-Ed

Straker
Posts: 657
Joined: Fri Jul 23, 2004 11:10 pm
Location: AB, Canada
Contact:

Post by Straker » Tue Sep 14, 2004 8:05 pm

yeah, what you described was more or less what i was guessing/asking, didn't think you were going to try to actually connect a pump right to a block. :P

hmm... was googling to find the weight of the csp750 and found this thread: http://forums.procooling.com/vbb/showthread.php?t=10040
cute.

for some reason C-Systems seems to have taken the specs offline, how much do they weigh exactly and are they intended to be mounted any way besides with the adhesive pad?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 14, 2004 8:12 pm

D-Tek Customs also sells the dual CSP-750 kit that includes a special bracket that not only connects the two pumps together, but also itself is used for mounting to the case, but that's not in any way helpful to the way I plan to suspend the pumps almost in the middle of nowhere within my case.

Straker
Posts: 657
Joined: Fri Jul 23, 2004 11:10 pm
Location: AB, Canada
Contact:

Post by Straker » Tue Sep 14, 2004 9:28 pm

I remembered that thing, but there aren't any screw/bolt holes right? I was going to say if they're really light, you could just unfold a coat hanger and use that, coming from a hole in the back or bottom of the case. Kind of hard with no holes or anything in the pumps though. :P

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Sep 14, 2004 9:31 pm

Unfortunately, no holes, no. That would've made life quite a bit easier, actually...

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sat Oct 09, 2004 9:28 pm

*Ding ding ding*...

Update time!

I'm still awaiting the new motherboard to come in (scheduled for Monday), and once I receive it, I'll run through Russ' arithmetic calibration procedure using the testbed video card and the XP-90 cooler before moving on with the liquid cooling antics.

However, I wanted to post up some pics so guys can get an idea what to expect...

Image
The new pumps!
The inlets are 1/2"OD barbed fittings, and the outlets are 3/8"OD barbed fittings.

Image
An overview of the case, PSU, radiators, shrouds and fans.
Modification was performed by GC Reliable in New Rochelle, NY.

Image
The new reservoir, from up front...


Image
...and its fittings.
This is a Typhoon dual half-height bay reservoir, initially branded by Primochill, but it seems Voyeur Mods has pretty much bought the rights and production facilities to this item. The outlets are 1/2"OD barbed fittings, and the inlets are 3/8"OD barbed fittings, and the two spare inlet openings are capped. All fittings, and also the fillport cap, utilize gaskets that Voyeur Mods supplies with this reservoir. I utilized teflon tape for additional safety.

Image
The GPU block radiator, mounted in the case's modified intake.


Image
Here it is from the inside.


Image
The CPU block radiator, mounted out back;
there's a gap from the case, to allow the side panel to close, and to allow the rad to breathe without case air pressure affecting it.

Image
This is the modified section;
it moves and widens the opening for the radiator, and allows the fittings to poke through directly under the PSU.

For pics of my new water block, Cathar's Storm G4, please click here.

I'm still debating whether I should swap the roles of the radiators and have the front intake radiator for the CPU and make the rear radiator for the GPU branch. Technically, it shouldn't matter, and I should go for whatever allows me the shortest tube run in general. I'll analyze it more fully when all the important parts are in place. It is also likely that I will be ditching the chipset block; with the position of the chipset on my new mainboard, the DFI LANParty UT NF3 250GB, it's virtually impossible, if not incredibly difficult and inefficient to install and run the tubing for the thing.

That's it from me for now, until I finish the calibration procedure and install everything into it and start playing around with it. :D

-Ed

DryFire
Posts: 1076
Joined: Sun May 25, 2003 8:29 am
Location: USA

Post by DryFire » Sun Oct 10, 2004 10:51 am

Ed you know i can't wait. That looks like it will be a killer machine.

I just got a job building someone a computer so it looks like I'l be able to get a storm myself, but i have to wait for the next batch :(

snutten
Posts: 341
Joined: Mon Nov 25, 2002 1:27 pm
Location: Sweden

Post by snutten » Sun Oct 24, 2004 7:43 am

1911user said it. This holds all the more true if you w/c for silence. The choice is big rads or powerful fans.
All this talk about flow, northbridge blocks and shit is only for overclockers. Not to mention the ongoing hunt for bigger, louder pumps. Unless it's totally suspended in the case it's bound to make more noise than any quiet fan.

Not to say anything about the nice work presented here, explicitly for overclocking.

Post Reply