First time watercooling

[cx323]

I decided that i want to try out watercooling (just my cpu) and i've come up with a list of what i'm looking at to fit in a budget that a little over $100.

CPU Waterblock-Swiftech MCW6002/6000
Radiator-MCR120
Pump-MCP350
T-Line
Tubing
Fan-?
Antifreeze/Hose Clamps

I will be trying to find most of the items used in the classifieds so that should keep prices down some. What changes should i make to this setup?

I also wanted to know if it was easy to transport a watercooled rig to lan's or if you had to dissasemble it or something?

Thanks for any help

[Otter]

Can you tell us a little more about your goals for this system? Do you want to maximize your OC, are you just looking for a cooling system that won't keep you up at night, or both?

[johan851]

That looks like a pretty solid entry level setup. If you started with an MCP650 instead of a 350, you'd have a little more room for future expansion. I would recommend doing that, as both pumps have about the same pressure, but the 650 generally allows for more flow at a given pressure.

In terms of a fan, it'll be a tradeoff between silence and performance. You'll need to pick one.

If everything is pretty solidly in your case, you shouldn't need to do anything to transport it to LANs. Just make sure nothing's going to fly around.

[batboy]

You need a fan to blow air across the radiator. I think the MCR120 usually comes with a 120mm fan.

I have a Swifty MCW6000. It's a decent well built block. The only difference between the 6000 (3/8" ID) and 6002 (1/2" ID) is size of hose barbs.

So, now you need to decide what size tubing. How much cooling you need and whether you have tight bends will help decide that factor.

My advice is don't skimp on the pump.

If you build the system with everything mounted on or in the case, it will be completely "portable". A little heavy, but you can lug it into a LAN. Add a side window and some nice lights... well, folks will be impressed.

[rogerdugans]

Looks like a pretty good setup to me.
Might be a bit of a trick to find all those choices and stay in your budget, but given time and some effort you should be able to get close, at least.

As for transporting an H2O LAN rig-
If you build it with transportation in mind, you should have zero trouble.
Making sure that ALL fittings have the tubing securely clamped is a big concern, but I see you have clamps on the list. :)
A good, secure T-Line is important too: if you have a solid plug of some sort clamped in the end of the line, you should be all set. If you make sure that the system is completely bled and topped off right up to the plug in the T-line you will have few worries about getting any air back into the system during transport.

My test after getting a LAN rig done and configured is this: unplug everything, pick it up and turn it over a few times. Hold it on its side, back, every which way.
Then I check for leaks. If something leaks, fix it and repeat until nothing leaks.
Once nothing leaks (and any leaks have thoroughly dried out!) I restart the pc and see if it still works. ;)

The treatment I give the thing is worse than anything it is likely to get while being transported, so if it can handle THAT, its all set.

I have carried LAN/transportable H2O rigs many times, including trips from Massachusetts to Indiana, Virginia and other, closer states.
Never had a problem, and never drained or disassembled anything.

[cx323]

I think i am going to look for a MCP650. Also, i have a fm121 lying around, would it work on the radiator if i want performance, or should i look at a delta?

Thanks for all the help

Edit- How does the tdx compare to the mcw6000?

I just read that you need at least a dual 120mm rad to use a mcp650, is this true?

[johan851]

I just read that you need at least a dual 120mm rad to use a mcp650, is this true?
Don't know where you read that...that's total hogwash. Doesn't even make sense. Why would you need at least a certain size of radiator for a given pump?

The TDX will perform a little bit better (maybe 1 or 2 C, depending on heatload) but will cause more pressure drop over the system. If I were you, I would get the MCW6002 just because of the quality build and low price. They're great blocks.

To maximize performance, I would suggest you get components with 1/2" barbs, then use 7/16" tubing over the whole loop. This combines the low flow restriction of larger barbs with the ease of plumbing of small tubing.

[cx323]

Don't know where you read that...that's total hogwash. Doesn't even make sense. Why would you need at least a certain size of radiator for a given pump?

The TDX will perform a little bit better (maybe 1 or 2 C, depending on heatload) but will cause more pressure drop over the system. If I were you, I would get the MCW6002 just because of the quality build and low price. They're great blocks.

To maximize performance, I would suggest you get components with 1/2" barbs, then use 7/16" tubing over the whole loop. This combines the low flow restriction of larger barbs with the ease of plumbing of small tubing.

I read it over at xs, someone said that that the pump put so much heat in the water you needed at least a dual rad.

Thanks for the help

[Otter]

Perhaps whoever said that was thinking of the heat generated by the pump. The MCP-650 isn't a power hog, but heat would be the reason to specify a minium rad for a pump.

Edit: Oops. I gotta remember to check for new posts when I'm multitasking.

[rogerdugans]

As far as needing a certain size rad for a certain pump......

There IS a relationship involved because of the heat the pump will put into the coolant system, but there are other variables as well.

Saying that any standard pump REQUIRES a certain size rad is....assuming many things. ;)
Saying that a big, hot-running pump will work much better with a larger radiator is likely to be true.
ANY cooling system will work better with a better radiator though, so this shouldn't be a huge surprise. But note that I said "better" not just "bigger."

To explain:
The cooling system essentially has 3 aspects (in the model I am currently speaking of, anyway):
heat input (water blocks and pump)
transfer medium (water)
heat output (radiator)

Note that there are TWO or more variables adding heat- every water block in the system and the pump.
A larger/stronger/hotter pump will generally add both more HEAT, which hurts performance and more flowrate, which helps performance.
A better radiator will increase the POTENTIAL cooling performance of the system.
But it can be tricky to figure out all the variables in advance....
IF the rad you have is "underworked" with one pump, a stronger, hotter one may improve performance thanks to the increased flow rate; the extra heat from the pump may not go beyond the rad's capabilities.
If your radiator is already pretty much maxed out, a better pump could conceivably result in higher temps: more heat load and not enough reserve cooling to dissipate it.

So this does get into some tricky territory, as far as advanced planning...

While I haven't used one, I would expect that any of the pumps commonly in use would work just fine with the old standard Chevette heater core radiator.
I know that with a chevette core I have used a number of different pumps and cpus and NONE of them so drastically poured heat into the system that it overloaded the cooling capabilty of the rad.

[cx323]

Here is the what i'm looking at now:

CPU Waterblock-Swiftech MCW6002-$20
Radiator-BI-$20
Pump-AquaXtreme 50Z-$65

I want to know if there is anything i should change without spending more money, and to make sure that a BIX will be enough for a dc chip. I know a heatcore would be better, but i don't have the room in my case since it is a somewhat small mid-tower.

Thanks for your help.

[Otter]

You can cool just about anything with any size radiator. If your rad and fan can handle 100W at 5C over ambient and you need to get rid of 200W, there's no problem. But your coolant will be 10C over ambient instead of 5.

If you're planning to overclock a dual core CPU, I think you'll want more than a 120x120 rad. But you can always back off on the overvoltage if it's getting too hot.

[johan851]

rogerdugans is correct, of course, especially with the "tricky territory" part. I said that it didn't make sense simply because there are so many little variables to think about.