first loop: build log and experiences

[Whatts]

started leak testing

Almost a week after starting the build I can finally start leak testing. This setup should make gravity do a big part of the bleeding, even before starting the pump:

The rad is down low (and was pre-filled):

 

[Whatts]

So after putting paper towels everywhere I'm ready to start filling 'er up:

When the res was nearly full I put a little piece of tube in the fill hole to stop it spilling over:

 

[Whatts]

Hmmm, beginners mistake: the silver coil has gone halfway down the 45 degree rotary adapter:

At least it's not going anywhere from there, so... Live and learn! Repeating the bumping/refilling a few times only took a minute, working out the smaller air bubbles by shaking the rad another 30 or so. And now after about 2 hours the system is running silent, there are almost no glugs or whooshes anymore. So silent in fact I just checked the pump to make sure it's still running No sign of a leak anywhere, but I'll keep it running for the day to be sure.

 

[bluezero5]

i keep my coil in my res for that reason.

 

[Whatts]

i keep my coil in my res for that reason.

Next rebuild perhaps. This mistake was a free one, the fittings on the other hand...

 

[m0r7if3r]

Actually, with it being on the outlet, that spot is fine for it. It will never go into the water path in the block and obstruct something, and it's not gonna get stuck in an eddy like it could in your rad.

 

[Conumdrum]

Yep, it'sll be fine there for a long time. I make my Killcoils a bit fat and push it into the tube, and then bend it a bit more so it's should say where it was.

 

[Whatts]

Up and running, first temps

I've still got to finish off a few details (so more pictures will follow later) but after about 6 hours of successful leak testing and bleeding I booted my system for the first time in a week. First thing I wanted to do was get some temperatures under load, no overclock so everything on stock speeds. I used GOM Video Converter to get 100% CPU usage (recoding a few Glades episodes) and HWinfo to measure core temps (first column is actual temp, second is minimum and third is maximum). Also shown in the screenshot is the Asus Fan Xpert PWM profile:

This screenshot was taken after 15 minutes under full load The temps might drop a little still when the Prolimatech TIM has cured (though that shouldn't be more than 1 to 1.5° C according to several reviews). Ambient temperature was 21.0° C, the fans were undervolted and running at about 45% (inaudible) and the pump just under 40% (inaudible).
Result!

 

[ATMINSIDE]

Very nice!

 

[Conumdrum]

Try 20% low and 50% high inthe fan prifile, you'll be amazed temps stay great and it's quiter.

I run at 20/50 I think.

 

[wagex]

Hmmm, beginners mistake: the silver coil has gone halfway down the 45 degree rotary adapter:

At least it's not going anywhere from there, so... Live and learn! Repeating the bumping/refilling a few times only took a minute, working out the smaller air bubbles by shaking the rad another 30 or so. And now after about 2 hours the system is running silent, there are almost no glugs or whooshes anymore. So silent in fact I just checked the pump to make sure it's still running No sign of a leak anywhere, but I'll keep it running for the day to be sure.

dont feel bad mine almost went into the pump

 

[TheHutchTTU]

For the MCP35X, 60-100% PWM is essentially 100% RPM as seen here.

Just FYI.

 

[Whatts]

It's almost midnight over here but one last post before I go to sleep

Very nice!

Thanks

Try 20% low and 50% high inthe fan prifile, you'll be amazed temps stay great and it's quiter

I'll try tomorrow, right now I just set a straight line to keep it simple.

dont feel bad mine almost went into the pump

Almost? Lucky escape, can't think of worse places where it could end up...

For the MCP35X, 60-100% PWM is essentially 100% RPM as seen here

Around 40% is where is gets audible, so anything under that is optimal. I won't have gotten anywhere near 60 yet, considering my temps haven't risen high enough.

I'll post a few more temps tomorrow, including no load, no fans (I was curious).

 

[bluezero5]

For the MCP35X, 60-100% PWM is essentially 100% RPM as seen here.

Just FYI.

General statement.

that is the Curve given your PWM offers full voltage without problems. Fact however is MOST mobo PWM doesn't offer full 12V, even my Rampage 4 extreme cannot do so, when all the headers are occupied, the PWM measures at 11.2V only and that is considered GOOD to a whole lot of mobo we experiment on. That is why Fan controller becomes important in some cases.

That Pump curve will OFTEN shift to the right side. I won't be surprise to see the pump at 100% only if the PWM is at 80%.

 

[Whatts]

Fact however is MOST mobo PWM doesn't offer full 12V

I don't have a multimeter so I don't know what my mobo puts out, but that sounds like a good reason why the pump would run at 100% as from 60% PWM.

Some more temperature screenshots, first fans on max:

Maybe 2 degrees C less than with fans on 45 to 50%, but at max the fans are obviously very audible.
Next up is with the fans at around 35%:

A small increase over half power but nothing dramatic.
What would happen if one of the fans would fail? Or 2? Temperature with only 1 fan (around 35%):

Interestingly, the difference is minimal compared to running 3 fans.
So out of curiosity I tried without fans without load (the small spikes on core 1 are because of HWinfo):

Seems I could run it like this if I keep a (very) light load, but I wouldn't be at ease with it.
I'll probably end up with the fans at the highest RPM where they are inaudible, as there is little to gain at higher RPMs except for noise.