Well, yes... I guess I'll install the PSU extensions and SSDs again... and hope for the best. I'll post some pics as soon as stuff is in.
Dont worry about the whole water debate. Ive used tap water in the past, hardly any difference. Its all a money making racket tbh and blown completely over the top. Some people use a bit of biocide as the biocide from the silver takes a short amount of time to build up. As long as you have one or the other it doesn't really matter.
A pH that is too low will slowly attack your metal components, eroding surfaces sooner. Put a couple drops of detergent and that would raise pH back up and increase surface tension of water. Better to have the coolant slightly basic than slightly acidic. Acids oxidize metals.
Well, I thought detergent actually decreases surface tension of water... that's why it works as a fly trap. Also, I don't know if it would act weird when in small areas like the CPU block for instance (it's fashioned in the way of the Supremacy). Well, I guess I'll put stuff together this evening, and firing it up for the first time in about two months. Pics will follow, and what do stupid people like me do once their done? Worry about how shiny and led stricken it will be inside :3eyes:
Well... putting all the power cables and other plugs in, the system does seem to at least start normal. Recognised the RAM, looks like it knows there are two graphics cards plugged in, but as I wanted to check I realised, the idle temperature of the CPU in BIOS is 39°C.... isn't that way too much for the pump running as well as the fans on the reads at full speed (I believe)? edit: Or does that run with 3.5GHz in BIOS? :3eyes:
Oh, okay hehe... looks like I'm getting 39/40°C as load temperature then. I'm not sure if that's a good temperature, but to me, coming from air cooling, any load temperture beneath 60°C is a good thing so far... I have to remember it's load temperature, with idle most likely being way lower (ambient temp 24°C)
40'c is a nice temp, my crusty i7 950 runs about the same at idle with a max of 60'c on the cores when it's been on for a few hours. My biggest surprise when I went with water was the GPU temps. I used to hit 90'c easily with my old 570 but under water it dropped to 50'c top's and typically ran in the 40's.
Well, 40 is good for 3.5GHz / 1.01v, but the temperature seriously rises once I get to 4GHz/1.2v, which happened after I changed to the XMP for my 3GHz RAM. Now I'm running 54/55°C under load, which still would be OK I guess... I want to get to 4.4GHz maximum clock, that's my goal, and I hope it works... already needed 0.2v for a 500MHz increase Also, oddly enough, I don't trust that mainboard's BIOS... with me just enabling some option disabled by default called something with boost, it pumped over 1.5v into my CPU and temperatures rose to above 90°C in the time it takes to reboot into BIOS again... as if they try to fry my chip :3eyes: I didn't get to any benchmarking or gaming yet, so I can't really say much about GPU temps... I'm afraid what happens to my CPU temps once the GPUs are under load... I see bad moon rising.
I would forget the xmp way to overclock, you use unneccessary voltage which produces a stronger current which gives off more heat. I would look into offset overclocking which will give you very low cpu usage and voltage on the desktop. When you run something demanded the power kicks in, so your overclocking while undervolting at the same time. That's how I have mine, I hear that its easier to do on x99 than x79.
I still have to read all the guides and posts on how to overclock x99 / Haswell-E, so basically, coming from a i7-950 it's a bit more complicated than pushing in a bit of voltage, setting the multi to 24, and toy with the bclk I just enabled it to see how it goes. I haven't been doing any work at the rig since friday, as I had to move my HTPC from win8.1 / LGA775 to win7 / LGA1366, so now the 8.1 license I have is free to update again. Either today or tomorrow I'll hook it up and let it upgrade, while reading about CPU / RAM overclocking. I'll touch the GPUs later. On the good side, it seems like both the 980s are recognised by the system, so I didn't break one during installation of the full cover block, great success
Somehow I will still have to toy with the system, once I carry it to where it belongs today... everything is running, so far the good news. But then again, I've observed some rather strange things I didn't notice until now. First, the CPU does not downclock. I will have to confirm this later today, but as it looks the CPU does not downclock. I guess it has to do with the XMP profile or some stupid auto overclock 'CPU upgrade' whatever function from the mainboard, but I was just downloading BF4 by origin, and the CPU was doing 3750MHz with 1.2v (HWinfo reported 3749.something and 1.198something). Seems odd though, as I didn't think that my CPU should run full multi / clock when downloading over WLAN with 1.4MB/s of data incoming... Second, I had 'load temperatures' (my CPU does not downclock so I take it the usual temperature of around 65°C is normal), and when I let the rig download over night I left HWinfo on too. Today morning I checked, and it stated a maximum temperature sometime during the night of the 77°C (core 4)! I'm not even running a big overclock or big voltage, and already I'm topped out temperature wise? If this is normal I'm quite disappointed... wouldn't have to go through all the trouble just to end up with air cooling temperatures. I don't know what was wrong, as when I came back it showed 'normal' temperatures... Also I noticed that the aquaero had a red led on it (close to the bottom right aqua bus connectors), which was constantly on, and a quick search on the net showed that it is a common problem, probably resolved by turning the rig off, as well as turning off the PSU / unplugging the PSU from the wall's power outlet. Very strange too, since I've never seen this light before, and the rig has been running a few hours now... Also, couldn't download or install one windows update, but that I don't worry about tbh, as the other 60 something worked without troubles. The last month or so I couldn't download any more updates for my old gaming rig (win7) at all. As a rather basic question, maybe directed at users of the aquaero fan controller, do you guys (or Vet ) plug in your pump's speed plug to the aquaero or the mainboard? As I've currently put it on the Rampage 5's CPU fan controller (PWM mode), and running it with the rather normal 'profile' in BIOS, just to make sure this thing is always on when the mainboard is. I've toyed with it and let it run with full speed, and boy that thing turned audible, and also, powerful... if you know my reservoir (pic here), I've currently got it filled to just below 'MAX'. But when the pump turns up to full speed (haven't noticed it during 'normal' use yet), the water level right from the grid near the intake drops about one or two millimeters, the part left from it rises about the same amount, and to the top most left, the water actually touches the res's ceiling... damn that looks like a lot of preasure! Also, I read about the AX1200's OCP and that mydog was advised to disable it, would I need to do that too? I don't plan on that high voltages or huge overclocks as he did (Haswell-E official thread, see here: LINK). Somehow I'm not sure I would need that much power, drawing 1300W or more out of the wall with 1 overclocked 5930K, two (currently stock) 980s, threee SSDs and the pump plus fans...
First of all turn everything back to stock and run again and monitor temps. I do not have my pump connected to my aquaero. I have the pump connected to a normal 4 pin molex power connector coming directly from the psu. As for ocp mode yes disable it. Load up the corsair link software and disable ocp by interview cking the boxes and then just leave it. The red light on the aquaero is something ive never come across. Although the fan controller has many features I only use it for low rpm fans and the onscreen software which is excellent. The thing is when you use an xmp profile you use more voltage than you actually need, same as a cpu quick and dirty overclock without tweaking. I would not recommend xmp oc nowadays. Once you get these niggles fixed we can look at overclocking on a negative/positive offset.
Thanks for the reply mate. I'm just curious, because my pump comes with the molex plug (which indeed I hooked up to the PSU), and additional, it also has a fan plug, but only with two pins there. Supposedly it's for receiving tacho speed, as indeed it does vary in speed depending on CPU temperature. I will need to unplug the front USB2.0 plugs altogether then, since one is used by the aquaero, the other now the PSU. So only the USB3.0 ports are connected up front, but I guess I can live with it. I'm also annoyed by all the changes that simple XMP setting for the RAM made... I could live with the 125Mhz strap, but certainly I've got the feeling it changes voltage and thus temperature too, which is indeed what is bothering me. Also, oddly enough, I need to toy with the BIOS because I think it also raised the multi to 30 instead of 24 and such things... I guess I will go with 2133MHz on the RAM first, then OC the CPU properly, test it for stability, then go for the RAM, test again, then go for the GPUs, better slow and safe than sorry.
Send this guy a PM on the Overclock.net forums....Jpmboy. http://www.overclock.net/u/279725/jpmboy He is a very helpful person and will go to great length to help you. He has his 5960k Overclocked using an offset and he knows what hes talking about. Ive Pm'ed him and explained to him that you need some help and will be in contact.
If you run high perf. power plan then it will run at max multi all the time, use balanced for idle.. Dont use offsets when you find your stable OC, use adaptive voltage. You can OC in XMP mode fine, just don't use auto bios OC (keep it "as it is" option), manually enter sync all cores and then input your desired multi. Also disable asus multicore enhance and don't use voltage - auto, use fixed voltage first then switch to adaptive with your stable voltage - total adaptive turbo voltage. Sometimes certain ram freq. adjusts blck to get higher ram freq., at least that's what I saw by few @ Haswell-E, usually anything over 2800mhz+. Also if you use win8.1, then when cpu is in idle it can sometimes do background maintenance and this can load 1 thread to the max, guess it was just that that made 70C+, but still kinda high imo for custom WC. Can you make few screens of AI Tweaker tab, also DIGI+. By cpu SVID try to stay bellow 1.85 -1.95V, this should be enough up to ~ 1.30 - 1.35v, also disable SVID (VCCIN - input voltage) and then manually enter e.g. 1.80v and go from there. And by cpu power management (should be bellow DIGI+ in AI tweaker) disable iVR fault management. Asus c-states (Advanced tab, next to AI tweaker) @ auto are usually off, and that will make fixed cpuVID all the time including cpuv, this will heat more too in idle if @ high performance power plan.. In my case 24-28C vs 34-38C. If you enable them incl. C1E, expect c-package (leave this auto /off), then also enable dynamic storage accelerator so iRST driver controls them later and balanced limits to C3, high perf. power plan to C0, otherwise it will run @ C7 by any power plan and this can sometimes cause instability issues by sudden higher cpu load @ higher cpu OC.
Oh boy, thanks hehe I'll go over to overclock.net and register there then, might come in handy after all (also tinkering with GPU BIOS and such). I haven't changed much about windows power plans, I guess it might still be in performance (thought they'd go for balanced by default). As for those 70+°C, I seriously don't know what went wrong... the pump was still running, and like I said, the next time I checked hwinfo it all was 'fine', except of course basically high temperatures... I wouldn't know because I never had a custom h2o system before, but 65°C at less than 4GHz and less than 1.3v pretty much looks like an average air setup... I hope to find out how to keep the temps down. Wow... I'm amazed at how you throw those abbreviations and numbers at me I'll have to check it out, I should have time for it either this evening, or tomorrow the latest. I'll post some screens too, as there's still the final 'rig mugshot' missing. It's kind of dark in that huge case, but it looks all nice and tidy at least.
Are you sure you don't have ahuge air pocket in one of your rads? Run a benchmark to get the temps up and feel the full length of the rads top to bottom. If there is a decrease in temp along the top half of the rads compared to the bottom and then you may have an air pocket.
Honestly I don't think so, but something's wrong... I've had the rig running upside down too, bled out quite some air that way, yet the temps certainly don't feel like what I expected somehow. Yet I don't hear any bubbling sounds or so when turning on the rig / pump... wouldn't an air pocket make noise? I mean, with stock 5930K settings I got a load temp of pretty much straight 40°C, I could live with that. But as soon as I hit that XMP things were starting to act weird.
Do you really want to use XMP? Normally XMP will change your bclk to 125 Mhz. If I recall correctly as soon as bclk strap is not around 100Mhz, the adaptive voltage is disabled in UEFI if you keep default voltages options. You can manually force it but I think using a 125Mhz base clock is definitely not a good choice. Is your pump speed controlled by the MB or is it working at a constant speed? If it can help, I can share my settings with you. On my new CPU I managed to have adaptive voltage working for both core and uncore. It never worked correctly on my previous CPU.
