I jsut activated XMP to see if the RAM was working properly, I was aware of the bclk increase, but not that it messes with voltage. I don't want to use it, but I don't know either if the RAM can run 3000MHz without the increase to the 125MHz strap. My pump currently is hooked up to the mainboard for PWM controll, power ofc directly to the PSU. I'm using the Rampage V's BIOS controlls for speeds, so it basically runs at 50% until about 50°C, then constantly increasing till 100% at 70°C. Not optimal though... Well sharing the stats might help me once I know what to do. The feeling I got is that I might still have to unhook everything and once again tilt, rotate, and flip to maybe get out some air, idk... still does not feel ready for truely toying with any settings as it gets way too hot for custom water with two 480 rads...
For the pump we mostly have the same settings on my side it is on manual with: 45c - 20% 55c - 35% 70c - 100% It is pretty quiet except under heavy stress tests. Do you hear any "glouglou" sound when the pump kicks in at power-up?
XMP profile gives way too much extra voltage that is needed for a specific oc. I tested this myself, I know its on X79 but still......
Well yes it does make low sounds of bubbles in the loop, but I actually stuck my head inside the case (as it is huge), and thought to have tracked it down to the fan controller's mini block (it is watercooled too). It as a rather quiet sound, so I wouldn't expect it to be a huge bubble or air pocket, but merely small bubbles that I thought to go away over time. I will plug in everything to start the rig now, and revert to stock settings in BIOS, so I will check again how the default behaves temperature wise. Thanks for all your suggestions and help guys, it's much appreciated!
Here are some BIOS shots, I reverted everything to default, I guess. I also set the mainboard BIOS's power settings to efficiency rather than performance, so the CPU downclocked. So, first things first: idle temperature seems to be fixed at 34/35°C, with about 24.5 as ambient, resulting in a delta t of about 10°C during idle / BIOS, with idle voltage too (check screenies right side 'hardware monitor'). With such temperature atm I can't really tell if there's big differences with the rad's to / bottom sides, but I'll check once I go back into windows and maybe run some Prime95?
I have about the same idle temp (room temp is also close) I only have one 360x60 rad in my loop but I think in idle it should not make any difference. On my side I kept performance profile. It is maybe stupid, but it garantees that boot will be done as fast as possible. Once windows is loaded, it takes around a minute before clocks are slowed down. Of course, balanced power profile must be selected in windows settings. On my side I'm using "standard" profile for "CPU power phase control". Not sure it is making any difference though Prime95 is a solution, but i find cinebench is even better. You can use a tiny soft like mouse clicker to emulate a click every second and run the test forever
you can leave everything alone in DIGI+ then. Nothing out of ordinary by advanced cpu settings either. Btw I was thinking mostly main Extreme tweaker menu settings :nerd: from my experience, it uses less VRM power, lower heat vs extreme, asus optimized should be optimal for that perfect balance, sometimes standard could underpower it and maybe glitch and hardrest or it needs more cpu current power %.. Now I'm keeping all at auto since last few bioses, it seems Asus made it more robust and now it uses correct setting. But then again it could act alittle different on HW-E
I've 'used' the rig a little yesterday evening, and noticed while copying files from one drive to the other, or over the network from the HTPC to the gaming rig and vice versa, when the system ramps up to turbo mode (3.7GHz, about 1.05 to 1.07v shown in hwinfo64), it does reach some 55 to 58°C. I still feel something is seriously wrong. 1.07v and 58°C on the hottest core does not work out for me... I have almost no headroom for overclocking the chip to anything above 4GHz. I'll try to toy with the system a little, first I'll set the pump to run at 100%, and then I will see how temperatures will behave with Cinebench, but I've got little hope. The rads don't even get warm enough so that I might feel a temperature difference between the bottom and lower parts...
This might not help you at all, but when I first built my current waterloop I had similar issues with my temps being higher than expected, I'd actually managed to put to much coolant into my res and the pressure was causing problems with the flow of coolant, since then I've always left about 1cm free of coolant in the top of my res and it's helped quite a bit. Obviously it's difficult to compare an old CPU like mine with newer chips but mine idles around 25/30c with 64c being highest temp I see during OCCT at 4.4Ghz with 1.35v and HT on. Also it might take a few heat cycles for the thermal paste to cure depending on what you used, that will drop your temps by a couple of degrees.
Well I've used Gelid GC Extreme as the paste, I guess it might need a few cycles, but the temperature difference seems to be quite high for just the TIM needing to settle and work out. I've taken care to stay below the 'max' mark on my res, but it's worth a try to get a bit of it out anyway. What I noticed is that when I turn the pump to 100% duty, the idle temp rises from delta t of 10°C to a whopping 70°C over about an hour of idle time, I guess that full power on the pump doesn't really help it at all. Will try to set a minimum duty cycle in BIOS, since the 'silent' setting described above showed better idle temp. I'll try to lower the duty until 65°C to 20% and see where that takes me, after a few seconds temperature already dropped 2°C. Also, as I have used the win 8.1 license on my old HTPC, now I have to reactivate it again, needing to be done over phone. Boy, that's annoying! That's the second time only I installed it, and had it over a year on my old rig, and already they think I'm using it on two although the other rig is not in use anymore... annoying m$ bs you get for actually buying a perfectly legal copy....
Maybe you put too much or too little? Or like you said its not at its full potential yet.. Noctua paste is really good too.
Well, hard to know, but I tried to keep it to the monoblock's specification, as in, putting neither too much nor too little on it. Also I've noticed, during small use again now, that the temp with 1 to 3 cores loaded seems to reach just about 40°C for 3.7GHz, I still got to run cinebench though to see how it works once every core is loaded.
If the rads are not warm all over when benching or maxxing out the machine then you have a huge air pocket in there. This would also explain why your temps are high as there is a small amount of water in there that is very very warm to touch?
Well I would need to use a proper bench loop, as cinebench is over quite fast, and temperature drops rather quick back to stock levels (fans on the rads 1400rpm, pump at 20%). At least I got SLI to work, well... I will try to run cinebench continuously, maybe run a bit Prime95 and see what happens. In case I got such huge air pockets in the rads... how should I get those out? I've already done pretty much everything I can to get the air out before I even hooked up everything. I guess that would mean draining the loop, and refilling it? Jpmboy was decent enough to send me a tutorial on pdf, and some hints and guides on RAM overclocking, but with currently reaching 70°C and slightly above during cinebench, I don't think I will start to overclock right now.
try x264 fhd benchmark http://downloads.guru3d.com/x264-FHD-Benchmark-v1.0.1-64bit-download-2825.html btw it runs in realtime priority, temp reading will lag in that time..
Getting air out of fixed mounted rads is more difficult than you think. All you can do is vary pump speed, tilt case, lift case,try pump speeds at different settings with case at different angles as well as turning pump off completely as when turned on will create a large amount of pressure etc. A huge air pocket will not disperse overtime and even if it does your reservoir level will become low enough that its possible the pump will suck all the air back through the loop again. As for jpmboy, what I would do I just send him some pictures of your bios, tell him you would like to hit around 4.3 to start with. He will tell you what setting to use for the offset. Then all you have to do is apply voltage to what i would think would be a positive offset.....easy once all other settings are in place. I would suggest some settings myself but I do not have an x99. Are you sure you got rid of all the white crap in the rads before fitting if there was any?
Thanks TJ, but to be honest, I don't think there's much sense in running benchmarks now. I downloaded the tool, but I ran a few minutes of Prime95 yesterday (MINUTES), and saw the CPU temperature rise to 80°C and above on some of the cores and turned it off. There's something fundamentally wrong with my build, so I guess I'll have to go back five steps again... I see... can't say I'm happy about all of that, but being this far in the doghouse the way back would be longer if not just as long as pushing through. I will see if there is a handy way to get the rad out of the case and turn it the way that it's ports are below the res, and at the topside of the rad. Well I guess so, I never have done anything with the rads and fighting flux, and it looked to be pretty okay. I flushed it with water afterwards until the rinse was clear again, and haven't seen any white flakes being pumped through the system and landing in the res. Jpm was kind enough to send me information and tell me how to start, but I feel there's no gain in applying them now, as you can see, my temperatures are bad beyond air cooling level.
Have you touched the cpu waterblock? If it is not hot, the problem is more at that level. Is it fasten correctly?
Do you have any pictures of how your loop is currently running? I saw the pictures a few pages back, has it changed since then?
