Seeking Thoughts from WC Veterans on the Watercool MO RA3 360 or 420

[BonzaiDuck]

There must be several folks -- I know Guskline's build is featured in a couple threads -- who use these external radiators of different kinds.

For the MO RA3's [I'll put in a link below] -- the manufacturer offers brackets fitting 120, 140, 180 -- possibly 200mm fans. For the 360 model, this would mean nine fans just for one side of the radiator. The "Pro" versions of these rads allow for fans on both sides of the cooler.

http://www.sidewindercomputers.com/wamo4xproblr.html

I'm wondering how you might power the fans:

-- An extension of parallel wiring from your PSU?
-- A separate power-supply or 12V AC-to-DC transformer?

I know you can get the latter from the local electronics warehouse, and you'd probably want to insert a switch. You'd simply buy the power-pack rated to exceed the total amperage.

Since the radiators are promoted to offer an option of totally passive cooling, this latter option wouldn't be much of an inconvenience -- unless you forget to switch it on during benchtests, gaming etc. It just wouldn't seem to be the most elegant solution.

I don't think I saw a wiring harness offered by Watercool as an accessory. Did I miss something?

Sidewinder offers these 4-fan splitters, ostensibly for 4x 120mm but it would serve just fine for any size 12V-3-pin fans.

http://www.sidewindercomputers.com/xsqu12fancas.html

 

[BrightCandle]

The best way I know of powering the fans is to group them into 3's are put them into a fan controller so you can control their speeds. You do need to make sure the channels are capable of enough AMPs but basically anything from Lamptron will do the job depending on your budget they have options. They don't have 9 channels however, nor do you want to manage fans individually so groups of 3 or so is about the best compromise between power per channel and easy management groups. Makes more sense for me as I have 10 fans split into 4, 3 and 3 on the different rads. You can also find fan combining electronics on watercooling sites as well or wire it yourself with a bit of sticky tape, some solder or an electric joint box.

I wouldn't ever try passive cooling, once you overheat a loop really bad things can happen. You are much better designing it so the fans are slow enough that you just don't care about leaving it on and running as normal and have that cope with all the heat you intend to put in a loop. 600 rpm is really quiet, 800rpm is pretty quiet as well if that is the goal and you can cool about 130Watts on a single 120mm fan on a decent radiator (I don't honestly know how the Mo compares, its quite thin so I suspect its more like 90 Watts per fan slot).

 

[BonzaiDuck]

The best way I know of powering the fans is to group them into 3's are put them into a fan controller so you can control their speeds. You do need to make sure the channels are capable of enough AMPs but basically anything from Lamptron will do the job depending on your budget they have options. They don't have 9 channels however, nor do you want to manage fans individually so groups of 3 or so is about the best compromise between power per channel and easy management groups. Makes more sense for me as I have 10 fans split into 4, 3 and 3 on the different rads. You can also find fan combining electronics on watercooling sites as well or wire it yourself with a bit of sticky tape, some solder or an electric joint box.

I wouldn't ever try passive cooling, once you overheat a loop really bad things can happen. You are much better designing it so the fans are slow enough that you just don't care about leaving it on and running as normal and have that cope with all the heat you intend to put in a loop. 600 rpm is really quiet, 800rpm is pretty quiet as well if that is the goal and you can cool about 130Watts on a single 120mm fan on a decent radiator (I don't honestly know how the Mo compares, its quite thin so I suspect its more like 90 Watts per fan slot).

Actually, I think either the 360 or 420 basic bare-bones radiator is close to 2.6" thick. The dimensions to the 360 are 415.5 x 383 x 65 mm -- more than half the width of a 120mm fan. I'm new to this -- no dispute -- but I think that's a pretty thick rad. Am I wrong?

Now . . . AigoMorla posted in a recent thread his thoughts about large fans with low static pressure. For each fan, the dead spots at the fan hub are bigger. On the other hand, doing a little trig arithmetic would tell you the dead-spot total area on an array of 9x 120mm fans (or 9x 140's -- for which I think there's an adapter for the MO-RA3 420). You can guess what I'm thinking about that.

If it were a matter of static pressure, one could buy the "Pro" model and install 8x 180mm units. Still -- one fan less than the 120mm array, and the dead-spots simply overlap.

Again, given the size of those monsters ( I mean the radiators), one might not be too frustrated by any desire to thermally control the speed. If, on the other hand, one could either get PWM fans of the bigger size (less likely) or control the fans from two 3-pin motherboard ports on the better mobos, you could use fans like the NZXT (180? or 200?) which boast a throughput of 166CFM at top-end (despite the static pressure rating). But they'd draw 1.4A per port if connected in parallel pairs.

Silverstone makes a 180mm offering:

http://www.performance-pcs.com/cata...=41689&zenid=8d0b00b8e533c43df7baa8b031c87058

It's billed as having good static pressure, which may mean the rating is relative to the fan size. But they also top-out at 900 RPM, and a rating of only 72 CFM. Probably doesn't matter much, though. If you purchased the "Pro" version of the radiator, you could put two in a diagonal configuration on one side, two on the other side configured at a right angle. You'd have two push, two pull -- only four fans.

The Silverstones are PWM fans. You could run them off a Swiftech 8W-SPL-PWM-ST or whatever it's called -- powered directly from the PSU and and controlled by the CPU_FAN or CPU_OPT_FAN port on the mobo.

However, at that speed -- ~900RPM -- why bother? Maybe -- maybe not. Easy either way . . .

UPDATE: Phobya offers 180mm fans, same RPM, CFM rating in the 50's. Same amperage or 0.21A. AND -- PWM.

UPDATE AGAIN: [Heh-Heh] The total dead-spot area for four NZXT 200mm fans is . . . wait-a-minute . . 4x 4.91 = ~20 sq.in. For 9x 120mm fans, it's ~ 16 sq.in. Dia for the NZXT = 2.5 in. and for either a 120 or 140 about 1.5 in.

Then consider all the dead spots from the fan corners. I think this is at least a "break-even." So that leaves the static-pressure as a factor. With that size radiator, for which the manufacturer boast "passive-cooling option," it leaves you "in bidnis -- definitely -- I think -- anyway.

AND AGAIN: This is entertaining:

http://www.youtube.com/watch?v=RGBXwtCQNvQ

I once installed one of those on my Honda Civic. It looks (and sounds) as though the fan is being tested for noise at different RPMs.

 

[BonzaiDuck]

I guess what I'd been seeking here was a "neat" solution to powering x number of fans. And like I said -- you could do it either way. Either a little "power-brick" plugged into a surge protector or UPS, or running a Molex to the computer-case exterior. It would need to be extended. For instance, if the fans were PWM, you could run an SATA power-plug out the case rear, hook up a Swiftech 8W-SPL. . . splitter, extend the wiring for -- say -- four fans, and wrap cables appropriately. You'd have to double back the PWM signal wire to the mobo. Or, leave the splitter inside the case. I say "four" because I was thinking of the 4x180 option.

With the right fans, you'd likely draw <= 12W. Nothing earthshattering to run off your PSU, I guess.

 

[aigomorla]

im very impartial about it.

1. i dont like things outside a case
2. i dont like things where i need to take apart to move outside the case.
3. i dont like compressor style heat exchangers as they are not efficient as hobby style radiators.

If i needed something that massive, I'd honestly rather bolt up 2 x 480's like i did in the old days which have way better cooling potential then a Mora does.

which was this wide:

and this long..

I called it my project to watercool the pacific ocean... err... no it was my TEC setups heat exchanger...

My TEC's

nah.... i dont want to go there again... wasnt efficient at all... so NTY!

 

[BonzaiDuck]

im very impartial about it.

1. i dont like things outside a case
2. i dont like things where i need to take apart to move outside the case.
3. i dont like compressor style heat exchangers as they are not efficient as hobby style radiators.

If i needed something that massive, I'd honestly rather bolt up 2 x 480's like i did in the old days which have way better cooling potential then a Mora does.

Did you have a power blackout today -- perhaps around 3PM PT?

ON TOPIC:

Yes to 1 and 2.

I'm looking to simplify the work as much as possible.

I think there's a way to consolidate case and MO RA3 by replacing the right side-panel with (a) the radiator and (b) 1/2" Lexan panel cut to the remainder of the exposed case -- if not that width then 3/8 or 1/4." 1/8" would take extra work. The Lexan part would fit the side-panel with thumbscrews.

The usual double-caster mod would extend the two frame pieces an extra 3 to 4" from the case. The radiator would sit on a third frame piece cut to case length/depth -- front to rear. The radiator would either be flush with the bottom right-angle of the case edge going front to rear, or would be cushioned there by a piece of rubber window-film molding or equivalent. If needed, there may be a way to further secure the radiator -- perhaps without use of screw fittings. The tubing would do a 180 turn at the back of the case, and enter the two rubber ports at the top-case-rear.

You could have quick-release fittings outside the case, but you might not need to use them when accessing the modified sidepanel(s) if the hoses are just long enough.

I would think you'd use pusher fans on the outside of the whole aseembly, pushing air into the case, and venting through what had once been the ideal intake ports. On the other hand, if that heats up the motherboard pan, you do the opposite and make sure the intake fans in the remaining ports add net pressurization to the case interior.

For your own grand design in the pictures, I'd try to do it with 12 fans while leaving the appearance and width identical to what you have there.

But I'm just looking for good cooling, fewest parts to put together, fewest steps, and a design that doesn't cramp your efforts to "access things." With the Lexan, I think you could add bling with a string of LEDs, but -- who cares?

It's likely I'd mod an old HAF 922. Put the sidepanel in storage. And keep the original caster assemblies for a new day.

 

[guskline]

aigomorla, when you refer to better cooling potential for 2 480s are you referring to the rev 3 MO RA3-420? I think you may be in error as to the cooling potential.

The revision 2 MO RAs were not as good but the most recent version was quite an improvement. The MO RA3-420 Pro is the equivalent of 3-420 rads and has the capacity for 18-140mm fans.

 

[BonzaiDuck]

aigomorla, when you refer to better cooling potential for 2 480s are you referring to the rev 3 MO RA3-420? I think you may be in error.

The revision 2 MO RAs were not as good but the most recent version was quite an improvement. The MO RA3-420 Pro is the equivalent of 3-420 rads and has the capacity for 18-140mm fans.

I'd have to go back and check, but there is a 360-Pro version with extra brackets to fit 4x 180mm fans.

For the 420, there is an accessory metal plate with four 180 to 200mm holes cut, and an array of mounting holes on the corner and the center dead-spot.

I assume that the proper mounting of the plate will create such a dead-spot in the center. But you're going to get that obstruction even if it's possible to mount the fans glued end-to-end and side-to-side. And as I'd shown, the fan-hub dead-spots for the 180-to-200mm fan array are only four square inches greater than with a 9x array of 120's, or 140's. I'm thinking that the corners of each fan added up for the 180's or 200's have a restrictive equivalence to the arrays of smaller fans. It might actually be less, but it could be computed with a fair degree of accuracy.

However, there's a logic to it, and I'd just go forward assuming that it's a mostly break-even situation.

That leaves the fans' static pressure. However, with a pressurized case, you might get reasonably good airflow even if the the radiator is mildly obstructed by the mobo panel.

There are accessory mounting brackets for case side-panels, but the side-panels must be flat. If they are not flat, you're likely to be working with a simple clear Lexan plate with fittings for these brackets. The brackets, which look like very long and flat cabinet handles, would leak air between the inner side of the radiator and the case-interior.

So that leaves a question: Should they be blocked off, if chances for pressurizing the inner case enough to just force more air through the remaining case obstructions and distribute reasonably across the radiator fins (?) I think they should. I think the same-size case intake fans (there are three possible places to mount them) could actually force more air through the fins and an array of 4 puller-fans on the opposite side of the radiator. The radiator would be an obstruction to begin with. Then the case-interior parts -- mobo tray, disk cages, etc. Since the brackets would place the radiator further from the mobo tray and other obstructions, the greater and more even the airflow.

I had been leaning toward the 360. The shipping weight on those things is something between 10 and 12 lbs. IF that's mostly the radiator and the accessories you'd use, it would require a sturdy side-panel mounting. Either sheet-steel or 1/4" Lexan -- to replace the original panel if it were not a flat surface. [Bulges stamped in the HAF 922]. But you'd have to cut the donut-hole, drill the holes, tap the case frame for 6/32 thumb-screws, . . .sanding with primer and paint for the sheet-steel. I think the Lexan makes more sense.

 

[guskline]

BonzaiDuck I also have a 922 case. I'm using it for my 8350 watercooled rig (actually was able to mod it to place a EK 480 rad on the top).

Mounting a MO Ra3 360 to the side panel will be a challenge AND restrict air flow. Obviously my MO RA3-420 is bigger so I left it free standing. My purchase included the pedestal feet (very nice). I bought extra tubing and 2 sets of KOOLANCE quick release couplings. This enables detachment of the MO RA from the main system relatively easily. I have BIG MO sitting about 18" center to center from the main case. Gives plenty of room for air flow.

 

[guskline]

aigomorla: Here is the only link I could find that described the difference between the MO RA2 and MO RA3 versions:

http://www.overclock.net/t/923448/u...0-26-supernova-mora-2-mora-3-phobya-nova-1080

I do agree that the flow rate on the MO RAs is lower ( I run D5s in series to address this). I just question whether or not your dual 480s would have a higher cooling capacity that a single MO RA3-420 Pro.

 

[BonzaiDuck]

BonzaiDuck I also have a 922 case. I'm using it for my 8350 watercooled rig (actually was able to mod it to place a EK 480 rad on the top).

Mounting a MO Ra3 360 to the side panel will be a challenge AND restrict air flow. Obviously my MO RA3-420 is bigger so I left it free standing. My purchase included the pedestal feet (very nice). I bought extra tubing and 2 sets of KOOLANCE quick release couplings. This enables detachment of the MO RA from the main system relatively easily. I have BIG MO sitting about 18" center to center from the main case. Gives plenty of room for air flow.

If the inner side of the radiator is set off from a hole in the case side panel of maybe 1/2" -- size of the handle-shaped accessory-brackets, it might reduce that problem. The orientation of the radiator itself on the case panel might also help. I don't know anything about fluid-dynamics or how to gauge the impact of case-interior restrictions, but I'd think it would boil down to a net area where flow takes place from the intake side. But you'd have to have good air intake from other case panels.

It would also be a Rube Goldberg that -- well -- simply looks like s***.

The EK-480 means 4x 120mm fans, and that's "just the way it is." I shouldn't have so much bias for fan-size. Now I'm curious to see what you did with the 922 and the EK-480 on top.

All the little design and inconvenience factors put a compromise to airflow, water-flow rate, radiator area and thickness up for compromise. For the trouble and aesthetics, I'd still be leaning toward the floor-standing radiator even if the case-mounting were made on the left side-panel -- resolving the matter of obstructions to airflow -- even if latter were only an acceptable compromise.

 

[aigomorla]

aigomorla, when you refer to better cooling potential for 2 480s are you referring to the rev 3 MO RA3-420? I think you may be in error as to the cooling potential.

The revision 2 MO RAs were not as good but the most recent version was quite an improvement. The MO RA3-420 Pro is the equivalent of 3-420 rads and has the capacity for 18-140mm fans.

this is why no other vendor will go back to compress style pipes, with the exception of aquacoolings tower.

aigomorla: Here is the only link I could find that described the difference between the MO RA2 and MO RA3 versions:

http://www.overclock.net/t/923448/u...0-26-supernova-mora-2-mora-3-phobya-nova-1080

I do agree that the flow rate on the MO RAs is lower ( I run D5s in series to address this). I just question whether or not your dual 480s would have a higher cooling capacity that a single MO RA3-420 Pro.

The phobya uses a flat tube which is different from the Mora.
Round tube is the same type of heat exchanger found in a air conditioning compressor.
They work good when its gasses which u need to condense, however if filled with liquid, they lack the efficiency a flat tube radiator will have.

Mora:

you see those round tubes... this isnt the case for the phobya:

It uses flat tubes... which id say the phobya is a much more efficient radiator.

However at those sizes, i guess it wont really matter, but one could get net the same effect using smaller rads in series instead of one very large one.

 

[BonzaiDuck]

It uses flat tubes... which id say the phobya is a much more efficient radiator.

However at those sizes, i guess it wont really matter, but one could get net the same effect using smaller rads in series instead of one very large one.

You might remember I'd mentioned the Phobya 200x200mm radiator in another thread. I think I may have also suggested a serial radiator configuration. But I notice in the sticky or other mentions implying that the serial approach "wasn't good for starters" or noobs to the water-cooling business. I think the caution derived from the number of connections or an increased likelihood (however small) of leaks. You can elaborate or correct if you wish.

Looking at the massive unit in your pics -- similar in size/dimension to the MO RA3 360 -- I ran some more web-searches.

http://www.frozencpu.com/products/2...n_2_-_Full_Copper_.html?tl=c95s1586b189#blank

http://www.performance-pcs.com/catalog/index.php?main_page=product_info&products_id=32968

I'd think this second "revision 2" item has half the cooling capacity of the 60mm version. I'd only guess that the 60mm unit uses flat tubes.

But with that, you'd wish for a case like the HAF X or 932, but with the motherboard closer to case-bottom and some alternate location for a PSU. Just guessing on that one.

All being said, one might put the 200x60 Phobya on its side and above a bottom-mounted PSU. You'd only wonder if it would clear a single graphics card, so what if you wanted SLI?

Figure you're looking for the right-sized vent (or "vent-potential") with 60mm+25mm+ [increment] of space that doesn't obstruct the motherboard or gfx-card placement.
UPDATE:
Maybe just to jumpstart more conversation, I've now discovered that the Phobya Xtreme 400 would require modding the upper part of the case -- maybe cutting into the 5.25" cage assembly. Some mods actually put the radiator on top of an Antec case, still butchering the 5.25" cage below and within.

 

[BonzaiDuck]

The clutter ruins the presentation. It's not the "bling" but the large caster-wheels, bottom-panel dual-140mm fan cage, and potential for further modding toward a larger radiator that means anything:

I gave the whole case and computer assembly to my brother. I'm not likely to get it back, and I'm not an Indian-giver, for want of an expression that isn't slightly "un-PC."

 

[aigomorla]

bonzai if your going to spend some funds on this setup, get a Mountain Mods, CYO, hori.

As you heard me toot many times, im not a big fan of offsized rads at all.
They are difficult to install, they lack the partnership with performance rated fans.

The only way i'll trust a radiator to do what it says it does is if it has the required fans.
If it doesn't have the required fans, then something with the smallest FPI (fins per inch) is typically signature for low rpm / low static fans.

The higher FPI's will require high static fans, otherwise they will not perform.

Lastly, I showed u that picture of the dead zone in the center of fans.
The larger the fan, the larger the dead zone unless the cone in the center is off spaced.

However if the cone in the center is small, then the fan is using a weak motor, and will suffer static from having to use blades which allow the motor to turn.

 

[BonzaiDuck]

bonzai if your going to spend some funds on this setup, get a Mountain Mods, CYO, hori.

As you heard me toot many times, im not a big fan of offsized rads at all.
They are difficult to install, they lack the partnership with performance rated fans.

The only way i'll trust a radiator to do what it says it does is if it has the required fans.
If it doesn't have the required fans, then something with the smallest FPI (fins per inch) is typically signature for low rpm / low static fans.

The higher FPI's will require high static fans, otherwise they will not perform.

Lastly, I showed u that picture of the dead zone in the center of fans.
The larger the fan, the larger the dead zone unless the cone in the center is off spaced.

However if the cone in the center is small, then the fan is using a weak motor, and will suffer static from having to use blades which allow the motor to turn.

I thought I'd posted the results. I'd overestimated the dead area for an NZXT ~200mm fan rated for 166 CFM. I used 2.5" diameter versus what seems more like 2.25". Then -- the diameter for the 120s or 140s is approximately 1.5" , For 9 120's versus 4 180s, the small fans are smaller in dead spot area by only about 4+ Sq. inches.

The NZXT tops out at about 1,350 +/- RPM, and can be controlled to as low as 300. I think the noise is mostly air-turbulence at the top end.

For a single fan, the dead zones between units of either size vary considerably. But we're comparing four fans versus nine fans. So "4+ square inches."

The remaining part of the exercise is to compute the dead areas at the corner of the fan shrouds. So again -- 4 versus 9 -- pretty close there too.

It's pretty much a break-even there. That leaves the static pressure, and the possibility that the case is pressurized and sealed except for intake and exhaust.

With the right-sized case, you could have four 180mm intakes and the four exhausts, and you might even fit pusher fans or above and below a top-mounted radiator.

And -- really -- I don't like large fans for aesthetic reasons. But they can be part of an effective airflow strategy.

Put another way. If it works with a heatpipe-cooler, it can be made to work for a custom-picked radiator.

 

[aigomorla]

I thought I'd posted the results. I'd overestimated the dead area for an NZXT ~200mm fan rated for 166 CFM. I used 2.5" diameter versus what seems more like 2.25". Then -- the diameter for the 120s or 140s is approximately 1.5" , For 9 120's versus 4 180s, the small fans are smaller in dead spot area by only about 4+ Sq. inches.

The NZXT tops out at about 1,350 +/- RPM, and can be controlled to as low as 300. I think the noise is mostly air-turbulence at the top end.

For a single fan, the dead zones between units of either size vary considerably. But we're comparing four fans versus nine fans. So "4+ square inches."

The remaining part of the exercise is to compute the dead areas at the corner of the fan shrouds. So again -- 4 versus 9 -- pretty close there too.

It's pretty much a break-even there. That leaves the static pressure, and the possibility that the case is pressurized and sealed except for intake and exhaust.

With the right-sized case, you could have four 180mm intakes and the four exhausts, and you might even fit pusher fans or above and below a top-mounted radiator.

And -- really -- I don't like large fans for aesthetic reasons. But they can be part of an effective airflow strategy.

Put another way. If it works with a heatpipe-cooler, it can be made to work for a custom-picked radiator.

dont get me started on NZXT fans... OMG...
The ones which came with the H440 were all Dog dookie.

I had to replace all of them to Yates.

The motors on NZXT fans are like... piss poor compared to even a Yate Loon.

 

[guskline]

aigomorla, Thank you for the explanation of round vs flat tubes. I see where the efficiency comes from.

Perhaps MO RA3 "makes up for it" by sheer size.

 

[BonzaiDuck]

aigomorla, Thank you for the explanation of round vs flat tubes. I see where the efficiency comes from.

Perhaps MO RA3 "makes up for it" by sheer size.

That, it probably does. Aigo mentioned a Phobya square cooler, much like the MO RA3. I found some smaller Phobya radiators that have a width of maybe 45mm. Some larger. But it looks like all that I window-shopped had flat tubes.

The fat lady hasn't sung yet about the "E" processors, but the prospects don't look all that good, even if you could keep the processor within 10C of room ambient.

So that's a major issue for me, but the fat lady hasn't sung. If I decide instead to use a Devils Canyon -- maybe 4790K -- then I might spring for a radiator that has cooling capacity of significantly more than the Nepton 280. A thicker radiator, maybe with a pair of 140mm Akasa Vipers. (Or -- two on each side, if it could be managed -- but it would be a "special effort" depending on the case.)

On the NZXT fan. They have a rifle bearing, but I think the MTBF is short. In addition to that particular spec, I think I burned one out. I'm guessing that a replacement would be the BitFenix Spectre Pro. Those are "rated" at about 144CFM. But if I don't use a huge radiator (as with 9x 120mm) -- no need for "huge" fans in smaller number. I think there are 200mm x 400mm radiators, or that Phobya might make one, but I'd have to look again.

Even so, the point of the NZXT reference was a measurement of the hub. The hub is either bigger or the same as any other 200mm fan. So the point of it was to get an idea of 4 dead spots versus 9. That would also apply to the 180mm fan offerings.

 

[guskline]

BonzaiDuck: I hear you about the fat lady singing.

My rig listed in my sig below is not my only rig. I have a 3770k @ 4.5 Ghz with the same mb and ram as you with 2 EVGA GTX670 FTWs in SLI. I also have a FX 8350 @4.6Ghz with an Asus Sabertooth 990FX (rev1) 16 GB ddr3 and it is custom watercooled and now uses my WC EVGA GTX780 Classified.

The rig in my sig is a classic Custom Watercooled "toy" (ie money pit).

Gaming wise my 3770k is it's equal using the same gpus. The 8350 is not bad but a little slower.

My point is that custom water cooling is expensive but fun. If you are a "tinkerer" jump on board with this warning. Be parepared to spend some serious startup $$.

 

[BonzaiDuck]

BonzaiDuck: I hear you about the fat lady singing.

My rig listed in my sig below is not my only rig. I have a 3770k @ 4.5 Ghz with the same mb and ram as you with 2 EVGA GTX670 FTWs in SLI. I also have a FX 8350 @4.6Ghz with an Asus Sabertooth 990FX (rev1) 16 GB ddr3 and it is custom watercooled and now uses my WC EVGA GTX780 Classified.

The rig in my sig is a classic Custom Watercooled "toy" (ie money pit).

Gaming wise my 3770k is it's equal using the same gpus. The 8350 is not bad but a little slower.

My point is that custom water cooling is expensive but fun. If you are a "tinkerer" jump on board with this warning. Be parepared to spend some serious startup $$.

Absolutely!! Positively!! But I've made a survey of the extra cash I'll need. Money isn't the main factor. the priority is to gather information so I won't be disappointed in my choices, and spend the money according to a "stocks and flows" discipline.

I have to consider what my preferences would be after the damn thing is built and running. I'm more inclined to keep the cooling solution within the computer case, unless there's a serious over-clocking benefit that results from doing otherwise.

Main thing -- I know what my options are on all counts. So far, anyway . . .

 

[guskline]

My HAF 932 limited the rad options

 

[BonzaiDuck]

My HAF 932 limited the rad options

Maybe I'm thinking about two things: aesthetics and the easiest case-mod options. Even without a "whatcha-ma-callit "brake" for bending sheet-metal, it might be very easy to fabricate a whole case side-panel without altering the case parts in ways that defeat their original function.

How much would a ~140-mm x ~280mm x 60mm weigh when full of water? Probably little more than the empty rad. It really wouldn't be hard to use a case providing a hinged radiator fitting with spring-loaded mounts. The radiator would sit just within the edge of the case-frame. The side-panel would have to accommodate the intake fans, which would just clear the radiator with its frame secured. You could seal the intake fans to mate the radiator with foam-board and acoustic foam rubber -- whatever.

So I'm wondering about the radiator-weight and the spring-loaded hinges. I'm thinking about the fan-cage of the CM Stacker midtower. . . . . 4x 140mm fan ports.

To access the motherboard, you'd remove the side-panel with fans, then unlatch and open the radiator assembly. You could even have a scissors-type hinge-reinforcement that would limit how far you open the "radiator-door." You have to choose a hose-length, so you might not want the door to open more than 60 to 90-degrees.

The uncertainty I have involves the weight and the spring fittings, but the item I'm thinking of has two latches with an interference fit to the drive cage toward the front case.

But there would need to be some modification or replacement of the plastic "inner-door" frame.

 

[BonzaiDuck]

Another Phobya -- on the big side . . .

http://www.xoxide.com/phobya-xtremenova1080-radiator.html

and maybe bigger?

http://www.performance-pcs.com/cata...oduct_info&cPath=59_457_458&products_id=32431

This one has more promise, but only for size. Narrower width -- less cooling capacity. Smaller dimensions -- less cooling capacity.

Single 180mm fan prospects. Can fit in to of a 922 or in the side panel, small enough to mount to the plastic frame I mentioned. Or, it could merely be fitted to the side panel, and the placement of an intake fan(s) might make use of the plastic frame -- as was originally intended by Cooler Master.

If mounted in case top, pusher fan would be more effective. Then you'd want to pressurize the case with two other fans > 180 -- in the usual lower front, and the side panel.

 

[Rubycon]

MORA "fan coil" style radiators and tank style (like used in PC applications and automotive radiators) are different designs entirely.

Fan coils can take pressure - a lot of pressure as a matter of fact.
Tank style cannot and can spring a leak with just a few bar.

Hobbyist WC pumps are quite wimpy and don't develop serious pressures.
However with all the tubing of the FC design, you will need a real pump (Worthington, Gorman Rupp) - ok that's a bit much but an Iwaki RD20/30/40 magnetic drive is not out of the question.

Those Laing pumps are OK but they are not designed for serious pressure applications. Their cut off head is quite low compared to a 24V RD30, for example.

And if you do design a high(er) pressure circuit, don't use compression fittings, SLI tubes, or boutique style clamps (cough, Lamptron) or you will be sorry.

To power these pumps and your fans you should always use a dedicated switching power supply of suitable wattage. Let your PC power supply do its intended job of powering the rig.