realistic temp diff on N/B only loop w/ 1/4 vrs 7/16

Been seriously thinking about running a second loop in my case for a NB and memory sinks only. Since my mem sinks are set for 1/4" ID tubing, I thought I might just put 1/4 fittings on the NB and run a single 120 rad for them. I know none of us guys want to say anything nice about "small tubes" , but I'd like a realistic analysis of temps on a small 1/4" loop like this. I'm obviously not going for world record wc temps but I'd like to keep that NB cooled and get more than 3.7ghz out of my system

What NB water block you goin to get, and do they even come with the "tiny 14" fittings?

I thought you had a 680i MB? considering the heat pipe is connected to the moffsets will you be water cooling the moffsets aswell?

Would you be better off getting a dominator fan for your ram.... and just adding the NB water block to your already existing loop with 1/2" tubes?

thx for the reply Myco,
well my ram sinks are already setup for wc ( or passive...OCZ 9200 flex) so since they are ready for liquid cooling AND I wanted to do the NB, I thought why not do them.
I planned to use the swiftech chipset block and have the 1/4 fittings extra left over from the Apogee GT I am using.
As far as the SB, I was going to passive cool that or use a small fan. EVGA makes a fan for the SB ...just want to eliminate some of the tubing that's gonna be in there.

I do have enuff 7/16 tubing left but reducing the line down for the ram sinks would unjustify it's use IMO...?

I don't know if it would be worth the effort. I think you are better off cooling them with air. 1/4 Tubing does seem kinda small.

It would make sense though to separate the NB and ram blocks because there are a lot more block out there that use small fittings. So keeping out of the main loop would make sense to not gimp your main loop.

--pak

JeffnWV said:

thx for the reply Myco,
well my ram sinks are already setup for wc ( or passive...OCZ 9200 flex) so since they are ready for liquid cooling AND I wanted to do the NB, I thought why not do them.
I planned to use the swiftech chipset block and would have to by the fittings extra (unless I lived in europe!)
As far as the moffsets, I was going to passive cool that or use a small fan...just want to eliminate some of the tubing that's gonna be in there.

I do have enuff 7/16 tubing left but reducing the line down for the ram sinks would unjustify it's use IMO...?

Click to expand...

I had a 1/4" WC setup, CPU, NB and GPU. GPU and NB temps were good (low-mid 30's), though the CPU temp would stay around 40 C. Honestly I think the CPU temp was do to bad contact between the IHS and core.
Dan

The NB and RAM don't make that much heat, hence 1/4" tubing should be enough.

Whatever size tubing you use, if you've only got one pump, you might have some trouble getting the right flow through each loop. A valve on the secondary loop might help with that.

Otter said:

The NB and RAM don't make that much heat, hence 1/4" tubing should be enough.

Whatever size tubing you use, if you've only got one pump, you might have some trouble getting the right flow through each loop. A valve on the secondary loop might help with that.

Click to expand...

People run two loops with 1 pump? I am going to guess you are going to split the loops with two T/Y fittings, but doesn't these defeat the purpose because the temperature will normalize anyway? Not to mention if one loops is more restrictive then the other, you are going to have definite flow issues like already mentioned.

--pak

Yes, parallel loops can be a hassle. A few people like this method for cooling things that don't need much cooling, though. The key is the light load on the secondary loop. They say that it's not too hard to find the point where the main loop is more or less unaffected and the secondary loop has adequate flow. If you're going to WC the RAM, it beats buying a second pump or letting the ram blocks take a big bite out of your CPU block performance.

pak said:

I am going to guess you are going to split the loops with two T/Y fittings, but doesn't these defeat the purpose because the temperature will normalize anyway?

Click to expand...

Eh? The purpose of a setup like this is not to create a temperature differential between the two loops, so no, it does not defeat the purpose. I suspect I'm missing your point, though.

For something like this, you'd use reducing T's instead of Y's. Because one branch needs only a fraction of the flow, a Y would just add unnecessary restriction to the main branch. Another way is to add extra barbs to the res and rad. This eliminates any added restriction in the main branch, but adds complexity.

never heard of it being done.. but a bypass setup would be what I would recomend for conrolling flow for a second loop with a single pump.

you would T off both sides of the main line and put a valve inbetween, closing the valve forces more water though the bypassed loop (out and in the T's).. and a gate valve has alot more controll vs. a ball valve.

I planned on using a second pump..a dedicated seperate loop. Just wanted to be sure that 1/4 lines would effectively cool the NB and 2 ram sinks.

No Y fittings or elbows expected