My first custom loop. Help me do a shiny!

[moonbogg]

The spring clamps that came with the barbs don't fit around the tube. They just don't open far enough even when sprung all the way open. I used zip ties on all fittings besides 3. I will get some on those last ones though. I didn't wrap the zip tie around the barbed section. I wrapped it behind the barbed section and pulled tight to create a tight neck area behind the barbs, so the zip tie is a smaller diameter than the barbed section and is trapped behind the teeth area. I'd like some legit clamps, but if these work then I'm fine with it. I can see the heat eventually softening and causing the tube to work loose, so I fully agree with clamps in light of your concerns.
And yeah, that SLI strap looks like a piece of trash. It works, but its ugly. I'll look at options.

 

[guskline]

The spring clamps that came with the barbs don't fit around the tube. They just don't open far enough even when sprung all the way open. I used zip ties on all fittings besides 3. I will get some on those last ones though. I didn't wrap the zip tie around the barbed section. I wrapped it behind the barbed section and pulled tight to create a tight neck area behind the barbs, so the zip tie is a smaller diameter than the barbed section and is trapped behind the teeth area. I'd like some legit clamps, but if these work then I'm fine with it. I can see the heat eventually softening and causing the tube to work loose, so I fully agree with clamps in light of your concerns.
And yeah, that SLI strap looks like a piece of trash. It works, but its ugly. I'll look at options.

You can buy hose clamps at Lowes or Home Depot.

 

[moonbogg]

You can buy hose clamps at Lowes or Home Depot.

I got some. I'll throw them on there. Don't want a leak. I can't believe how quiet this thing is. It basically sounds like a regular PC case with all passive cooling inside. All I hear are the case fans and they are giant things and are quiet. The old case had a jet engine H80 which was just ridiculous how hard it had to work to maintain 65c during a game. Plus the GPU blowers and that was one loud rig. I don't even care about OCing very much to be honest. Everything hauls anyways. I'm just loving the quiet, but of course, I'll OC some more.
Just did a few cinebench runs. If I recall that used to break 70c but maxes at 55 on hottest core now. I think I got some headroom to take advantage of.

 

[guskline]

moonbogg, impressive. Have you cranked those twin GTX980TIs up yet?:awe::awe::awe:

 

[moonbogg]

moonbogg, impressive. Have you cranked those twin GTX980TIs up yet?:awe::awe::awe:

Absolutely ridiculous and I got a little more I can squeeze out I'm sure. I am feeling kind of confused here because I don't recall my CPU being this easy to OC at 4.6. Temps hardly rose at all from 4.3 and it only took an offset voltage of .020 to get here stable enough for these runs. Voltage peaks at 1.288-1.29, which is barely above the 1.28 it used at 4.3. With less load on the CPU voltage hits right around 1.3 which is what I was expecting.
I noticed for some reason my CPU physics score dropped a little at 4.3 from what it had before. It seems the higher I clock the GPU's, the CPU scores a little less on the physics test. Cinebench R15 reports 1160 for the CPU at 4.6, so maybe you can tell me if performance seems correct here.

Max temps during these tests:

CPU @ 4.6Ghz
Coldest core - 56c
Hottest core - 62c

GPU's @ 1477 core, 3900ram, so 7800 effective with no throttling
Temps didn't break 42c (LOL)
I added 31mv of voltage to GPUs because firestrike failed to load before I tried the voltage bump. The bump appears to have fixed it.

Firestrike Standard

Firestrike Extreme

Do those scores look good?

I'll also say that Crysis 3 maxed with 2x TXAA hauls ass. The game just hauls absolute ass, and that's my scientific analysis. In the most intense spot I found, stock GPUs have about 60fps and with an OC a little less than this last one, it was about 80 in that same spot.

 

[aigomorla]

pay note to where i used zipties.. it has to be in the middle of the barb so it holds even if it is pulled back.

 

[moonbogg]

pay note to where i used zipties.. it has to be in the middle of the barb so it holds even if it is pulled back.

I see. I will add hose clamps everywhere I can fit them, and zip tie like you show here everywhere else. Thanks again. This water cooling stuff is grand. Love it.

 

[YBS1]

I'll get back to you later on when I'm home. Cinebench seems low for 4.6 but I'm going entirely off memory, and could be way off base.

N/M found the thread on mobile, looks like it's right where it should be.

 

[YBS1]

Absolutely ridiculous and I got a little more I can squeeze out I'm sure. I am feeling kind of confused here because I don't recall my CPU being this easy to OC at 4.6. Temps hardly rose at all from 4.3 and it only took an offset voltage of .020 to get here stable enough for these runs. Voltage peaks at 1.288-1.29, which is barely above the 1.28 it used at 4.3. With less load on the CPU voltage hits right around 1.3 which is what I was expecting.
I noticed for some reason my CPU physics score dropped a little at 4.3 from what it had before. It seems the higher I clock the GPU's, the CPU scores a little less on the physics test. Cinebench R15 reports 1160 for the CPU at 4.6, so maybe you can tell me if performance seems correct here.

Max temps during these tests:

CPU @ 4.6Ghz
Coldest core - 56c
Hottest core - 62c

You might be able to get closer to 5GHz than you thought you could get. Not a bad voltage at all for 4.6GHz. I'm wanting to say my 4.8 voltage is ~1.40 and my 5.0 voltage is in the 1.5-1.52 range from memory, I do know that I run my daily 4.6 at 1.34 (1.34 being peak voltage, it's variable, drops down below a volt when idle) but I don't recall if I settled on that because it was the minimum amount required or if I added a little to it for some stability margin. Chip might be better than you thought....:thumbsup:

 

[guskline]

Great job moonbogg. I ran my 3930k at 4.6Ghz 24/7

 

[moonbogg]

I'm wondering how gaming causes the hottest core to hit 62-64c, but running IBT at maximum doesn't make it go any higher. Its like the CPU has a min temp that it hits, but stressing it more doesn't make it get hotter because the water can absorb tons of additional heat or something.
Its odd that the CPU runs hotter than the GPUs, but after a certain temp it seems the water absorbs any additional heat being dissipated. There is some thermal dynamic stuff happening that I don't understand. There is no way in hell my H80 could have kept this chip that cool during an IBT maximum stress, let alone at 4.6ghz.
Initially I was disappointed with the CPU temp maxing in the 60's on the hottest core, but I'm OK with it I think because it STAYS that temp no matter how hard it gets stressed. That's very odd and very different from the behavior of my previous cooler. The H80 definitely became overwhelmed quickly with increased CPU stress, and the water loop seems to just keep absorbing more and more heat, preventing the CPU from going over a certain temp.

So hottest core during IBT maximum is 64 and coolest core is 53. That's the highest temps those cores have reached so the delta is about 10c between hottest and coldest core with the rest of the cores being somewhere in between.

 

[YBS1]

You could try a remount, on some blocks even the orientation can have an effect.

 

[guskline]

moonbogg: I would not remount. Give your system time to adjust and maybe tweek a little but seriously, 64C MAX core temp on a 3930k @ 4.6Ghz is phenomenal. You are dumping the heat from 2 GTX 980TIs and the heat from the 6 core 3930k @4.6 Ghz into a single loop. You have 2-360 PRIMO rads doing their job (glad you got the UT 60s!) and the max core temp is 64? Lordy, for most people to be so lucky!

BTW, gaming, if done for a period of time is stressing BOTH the cpu and the gpus so you are dumping a lot more heat into the loop than IBT which is only stressing the cpu.

 

[YBS1]

I suggested he could try it due to his concern over the variance in hottest to coolest core, not absolute heat. Some variance will occur in just about all chips though.

I know where he is coming from, I'm irked by more than a 5C spread too whether I can do anything about it or not.

 

[guskline]

I suggested he could try it due to his concern over the variance in hottest to coolest core, not absolute heat. Some variance will occur in just about all chips though.

I know where he is coming from, I'm irked by more than a 5C spread too whether I can do anything about it or not.

YBS1 I hear you. I have a spread between my hottest and coolest cores on my 5960x and chock it up to placement of the cores on the die. I have 2 cores higher and 2 cores lower than the norm but all running within acceptable temps. I'll do a long run of Intel Tuning software and monitor the variance and report back.

 

[moonbogg]

So, I did an experiment. I cranked the fans and compared the temps to when the fans are on medium. It was about 2 deg cooler. Then, I disabled SLI and used 1 GPU. The temps on the CPU dropped about 5C on average. This can only mean one thing. I need another rad! Maybe not right now, but soon I expect to buy a third rad and bolt that sucker on the back of the case.
My only real question here is, will a single D5 be enough for all that loop I'll have going on? 2 GPU's, a CPU and three 360 rads? The rads are pretty open as far as flow is concerned, at least I would think. I'd expect the blocks to cause the real flow restriction.

 

[MrTeal]

So, I did an experiment. I cranked the fans and compared the temps to when the fans are on medium. It was about 2 deg cooler. Then, I disabled SLI and used 1 GPU. The temps on the CPU dropped about 5C on average. This can only mean one thing. I need another rad! Maybe not right now, but soon I expect to buy a third rad and bolt that sucker on the back of the case.
My only real question here is, will a single D5 be enough for all that loop I'll have going on? 2 GPU's, a CPU and three 360 rads? The rads are pretty open as far as flow is concerned, at least I would think. I'd expect the blocks to cause the real flow restriction.

You should be fine. xtremerigs has an awesome 360mm rad roundup that shows flow restriction at different flow rates.
http://www.xtremerigs.net/2015/02/11/radiator-round-2015/3/

tl;dr - Most rads has <0.5PSI of drop @ 1GPM. Your UT60s have very little drop though. Most CPU blocks are between 0.5 and 1 PSI at the same flow rate.

 

[moonbogg]

You should be fine. xtremerigs has an awesome 360mm rad roundup that shows flow restriction at different flow rates.
http://www.xtremerigs.net/2015/02/11/radiator-round-2015/3/

tl;dr - Most rads has <0.5PSI of drop @ 1GPM. Your UT60s have very little drop though. Most CPU blocks are between 0.5 and 1 PSI at the same flow rate.

Oh wow, .06 is low. I wonder what effect the added weight and mass of the extra water has on the pump. Probably marginal. Thanks for the link and info. Now I just need to decide for myself if its worth the hassle and cost. I feel myself being pulled to get MOAR RAD.

 

[YBS1]

It's difficult to say if the additional D5 would give you any measurable thermal reductions due to the increased flow and head pressure. It would give you a margin of safety though due to the redundancy. We both know this stuff is our toys, it's really all up to if the ~$100 dollars bothers you.

Meanwhile I sit here with a pedestal, two additional 360's, another pump, and eight more fans still awaiting the stupid videocards I want, at this point I feel like Pascal will get here first. Maybe I should let you borrow the stuff, you'd make better use of it than me. That's why I mentioned them as toys, I have this stuff sitting here I can't currently get use out of and it's given me the itch to "assemble". I've been refreshing NewEgg all day checking on 6700K's, not sure if I'm hoping they will be in stock or relieved they aren't....The $^&# I need a 6700K for??? :hmm:

 

[guskline]

moonbogg, MOAR RADS HAHAHA! Just got done assembling an Asus Maximus VII Hero mb with a I7-4790k with an EK Supremacy MX block, 16G DDR3-2133, 2 R9 290s with EK gpu blocks and my trusty Watercool MO-RA3-420. I bought a Swiftech MCP35x pump with attached reservoir and all runs well.

 

[PhIlLy ChEeSe]

Nice work man! Love the case, nice small loop! You have a 1440 display, or multi displays?

 

[moonbogg]

The $^&# I need a 6700K for??? :hmm:

Because you must shove a plate of copper over it, add water and see how fast it will go? SOUNDS LIKE A GOOD REASON TO ME.

moonbogg, MOAR RADS HAHAHA! Just got done assembling an Asus Maximus VII Hero mb with a I7-4790k with an EK Supremacy MX block, 16G DDR3-1600, 2 R9 290s with EK gpu blocks and my trusty Watercool MO-RA3-420. I bought a Swiftech MCP35x pump with attached reservoir and all runs well.

Lol, yes indeed. Moar RAD is calling me already. That rig you built sounds pretty damn fast. I'm curious how the 4790K does on water.

Nice work man! Love the case, nice small loop! You have a 1440 display, or multi displays?

Not sure if you are asking me, but if so, thanks man! I use a single 1440@144hz display. Also, what do you mean by small loop?

 

[Sabrewings]

I'm curious how the 4790K does on water.

Mine will kiss 4.9GHz.

 

[guskline]

Mine will kiss 4.9GHz.

That's fast. I used the 4.8ghz setting in the BIOS and BOOM ran well. I'm using the new EK Supremacy MX waterblock which is not quite as good as your EK EVO block but is still a good block. I have the EVO block on my 5960x rig. I was going to sell the MO-RA3-420 external radiator but just couldn't part with it. I had a faulty Asus Z68mb (memory module problem) so I bought an Asus Z97 Maximus VII Hero mb plus a 4790k cpu (sold the 3770k chip which is a solid chip).

I found a good deal on a Swiftech MCP 35X pump reservoir combo which has tremendous flowrate for such a small pump. I also have the 2 Waterblocks (EK) for my 2 R9 290s so they are back in action in the 4790k rig. The MO-RA3-420 (has 9 XSPC 140mm fans on one side) gives me plenty of rad capacity.

I suspect it might run at 4.9Ghz but no need. Right now I used the OC wizard in the Bios and am running at 45x102 for an effective 4.6ghz rate and Very cool. The Devil's canyon is a nice cpu.
UPDATE: Settled on 4.7Ghz (47x100) Passed all Asus RealBenchmark tests with DDR2400 ram set at 2133.

 

[moonbogg]

For crying out loud guys, are you going to stop me or something? You helped create a RAD addicted monster here and now you're just going to let me run loose and fill my house with RADS?

No, but serious, I'm considering a third RAD. The only thing really holding me back (besides spending more cash of course) is that while a 3rd RAD might help now, it might go to waste after a future upgrade to lower power components. I'll flip a coin a few times for a best 2 out of 3 to decide if I buy another rad or not.
It would be another UT 360 RAD, a mounting bracket for the back of the case, and some fan cable splitters for another 3 fans. Also, more tube and a couple fittings. The loop would go from the GPU's to the new rad and then back to the CPU. That would keep the tubing lengths nice and short.