17 hours ago, martin-w said:

 

 

One thing to mention though. Something which often fools people. Any water cooling loop, and that includes AIO's, takes a while to heat up to full temp. It can take as long as 30 - 40 minutes for the loop to reach full temp. So strictly speaking any test on a water cooling loop should be conducted for at least 40 minutes before results are recorded.

The other point is ambient temp. Should be recorded for both coolers.

The D15 results at 4.9 you may have got mixed up, as they seem to be lower temps than 4.8. Great result though, very impressive. 

Hi Martin.

The tests were done for 20mins each, even the D15.  That's when the temps appear to stabilize on my rig. I guess because its only a small loop with a much bigger cooling capacity?

As for mixing up the temps...  I'm not at all surprised.  Should have gone to bed much sooner;-)  Not been so well and spent all day in bed yesterday:-(

Do you know which image servers still allow for external links?  Can't post directly here can we?

2 hours ago, Dougal said:

 

Do you know which image servers still allow for external links?  Can't post directly here can we?

 

To be honest I have no idea. Not posted images here before believe it or not.

 

This help...

 

 

 

14 minutes ago, martin-w said:

 

To be honest I have no idea. Not posted images here before believe it or not.

 

This help...

 

 

 

lol I found that just before you posted.  Still awaiting email verification from the listed link - doesn't seem to work. Guess i ought to find out how my own webspace hosting works too;-)

Overclock @ 5.0 GHz – vcore: 1.345v (Manual OC no XMP)

                                        Noctua DH-15 Air Cooler:     vs    Custom Water Loop Cooling:

                                          cpu0    cpu1    cpu2    cpu3         cpu0    cpu1    cpu2    cpu3

Idle min temps:                   23c    25c    22c    23c                22c      23c      21c      22c
IntelBurnTest load temps:   56c    72c    58c    66c                53c      69c      56c      67c
CiniBench max temps:        56c    57c    54c    66c                53c      54c      51c      64c

==============================================================================
 

Not much in it 5GHz manual, but always improves using XMP at 5

Lets see if this works...

pic1

pic2

pic3

pic4

Note the home made 'anti-sag' device for the graphics card.  Its net curtain wire, blacked with marker pen;-)

I've also 'hot glued' on casters.  With a stuffed back, its too damn heavy to keep lifting in and out!

 

Yep, see them Dougal. All working fine. 

I know what you mean about Heavy! Not good for dodgy backs.

Strange there such a small difference at 5 GHz. Usually that's when custom loops shine compared with air, at higher frequency. 

 

Is there a fan variable going on there? You should try it with fixed fan RPM.