Custom Water Cooling Mental Temperatures

[Sneakyfox169]

Hey everyone, building a new system on the x99 platform with a 5820k. For it I've gotten a custom loop in a 760t, a 360 30 on the top in push/pull and a 280 60 in push/pull in the front. However for some reason my loop seems to have no effect at all. the 5820k is OC to 4.3 ghz on 1.282 at the moment and it idles at 50-55 degrees celcius. I tried prime 95, and in the first second it went up straight to 100 degrees (I stopped it immediatly of course). So here I am, unable to understand what it going on. I have a alphacool 5.25 inch bay res with a Laing DDC-3.2, no matter how fast I make it go (or how slow) it seems to have very little impact if any on the temps, but front the way the water moves in the res I clearly don't have restriction problems... I'm using a EK Supremacy Evo by the way, I've had to put it upside down for the loop (the instructions say the inlet has to be on the hole in the center. So yeah, if someone's encountered similar problems or knows what's going on if you could let me know it'd be awesome cus I'm lost for ideas. Tried tilting the pc in every directions to make sure I got all the air out, shook it and no improvements...

[sissj]

I know its a stupid question but did you apply thermal paste?

[Sneakyfox169]

x) yeah, and there are no stupid questions at this point... 

[Enderman]

I know its a stupid question but did you apply thermal paste?

Or leave the plastic protective film on the CPU waterblock?

[theninja35]

Have you checked for leaks and made sure the water is making it all the way through?

 

Also you must be like hulk if you can just lift and shake your computer around.

[Sneakyfox169]

plastic cover is off, I'm wondering if I did something wrong when I changed the jets for the waterblock, I've put the "optimized" jets for 2011-3 according to the manual that came with it, but I checked after these crazy temps and I nothing jumped up to me and said, YOU MORON YOU PUT THAT WRONG, it was like in the manual... There are no leaks at least, used compression fittings everywhere and no leaks so there's at least that, but any sort of load and the temps rocket high up immediately straight up. 

 

EDIT: By the way it's quite hot where I live right now probably around 30-32 in my room, still I don't think it should account for such temps when I find in reviews that a h110i gt has temps of 64 under load with a higher clocked cpu...

[sissj]

plastic cover is off, I'm wondering if I did something wrong when I changed the jets for the waterblock, I've put the "optimized" jets for 2011-3 according to the manual that came with it, but I checked after these crazy temps and I nothing jumped up to me and said, YOU MORON YOU PUT THAT WRONG, it was like in the manual... There are no leaks at least, used compression fittings everywhere and no leaks so there's at least that, but any sort of load and the temps rocket high up immediately straight up. 

 

EDIT: By the way it's quite hot where I live right now probably around 30-32 in my room, still I don't think it should account for such temps when I find in reviews that a h110i gt has temps of 64 under load with a higher clocked cpu...

Its 25 Degrees Celsius in my room right now and with my Corsair H100i I'm hitting 30

[kennethfage]

maybe some air in the loop?

[Sneakyfox169]

I've thought about air, which is why I shook my pc as much possible in every direction, but I couldn't see any air moving in the hoses so, either it's some sneaky air, or it's hunkered down and doesn't want to move or... there is none. Is there such a thing as doa rad that don't transfer the heat?... 

[Sneakyfox169]

Well it's quite possible I noobd out, I took the block off, cleaned it off, and put much less thermal paste this time and lo and behold much better temps, idle 38-40 under load 65-70, still not quite good enough but at least I can run programs without it jumping to a million degrees. I used to use a hyper 212 evo from CM and that thing would have a heart attack if you mentioned a flat base surface where you can see yourself in the reflection, so you had to put more thermal paste than on a perfectly flat surface. So being used to that I put too much. Tomorrow I'll try putting even less on and see what it gives, my apologies for being such a noobie

[Sneakyfox169]

Well actually I lied, I had turned down the max temp of realtemp to 90 because I'm really not confy having the idea of t max being 100, even 90 but I didn't know that that meant that temps readings were 15 degrees lower than reality. SO here I am, still with unexplainable horrible temperatures which I have no clue why. On the up side I think I got a decent overclocker, from the internet it seems that lots of people are stuck at 4.4 ghz on 1.3v and higher, I'm running 1.285 at 4.5 ghz and it's stable and I could possibly even lower the voltage as I just put it there to see what would happen.

[Sneakyfox169]

I've taken the block apart again to see inside and again nothing seemed out of place, I took pics which I'll post later, I trying out running the block without the plastic thing that separates the water, I'll see what it gives...

[Sneakyfox169]

could my fans really be so slow they're the problems ? I mean yeah they're slow but still I'm in push pull with 64.8 cfm @ 1.21 mm H20 for the 140's and 54.8 cfm @ 1.48 mm H20 for the 120's...

[Sneakyfox169]

Things go from bad to worse, as some people say you have to flush even new rads, I figured at this point I didn't have anything to lose, but guess what ? I was wrong, while rinsing out the the rads, some water got on the pump which of course is not water proof, and guess what, I may now have a dead pump, awesome

[0x1e]

the pump should be fine if you let it dry out.

Once that's done put the loop back together and run the system at idle for a while.

Do you feel any heat coming off the rads? Turn up the fans if you need to so you can feel air moving.

[Sneakyfox169]

SO took the pump apart and actually it was dry. That said it had smoked earlier in the day. However it had kept working. I took it apart and a component on the pcb is black so clearly it fried and like I said the pcb area was dry, so water was no the cause. I'm wondering if I got a faulty unit or something, as it's my first time with a CWL I don't know much about pumps, I've sent an email to the retailer from which I bought it describing my issue. We'll see how it goes. Maybe the pump was always faulty and that's why my temps were so bad... Well one can always dream, we'll see how it goes, while extremely frustrating and aggravating at least it didn't fry anything else, so even if I have to buy another one (which would really take the whole build experience down some more) once I get a new pump either by the warranty or buying a new one, I should be able to fire it up again. At least it should let me know whether the pump was what was wrong with my loop or not. 

[TDP_Equinox]

Hey everyone, building a new system on the x99 platform with a 5820k. For it I've gotten a custom loop in a 760t, a 360 30 on the top in push/pull and a 280 60 in push/pull in the front. However for some reason my loop seems to have no effect at all. the 5820k is OC to 4.3 ghz on 1.282 at the moment and it idles at 50-55 degrees celcius. I tried prime 95, and in the first second it went up straight to 100 degrees (I stopped it immediatly of course). So here I am, unable to understand what it going on. I have a alphacool 5.25 inch bay res with a Laing DDC-3.2, no matter how fast I make it go (or how slow) it seems to have very little impact if any on the temps, but front the way the water moves in the res I clearly don't have restriction problems... I'm using a EK Supremacy Evo by the way, I've had to put it upside down for the loop (the instructions say the inlet has to be on the hole in the center. So yeah, if someone's encountered similar problems or knows what's going on if you could let me know it'd be awesome cus I'm lost for ideas. Tried tilting the pc in every directions to make sure I got all the air out, shook it and no improvements...

A little off topic, but where did you decide to put the res/pump in the case? Pics?

OT: Don't use prime95 on your CPU; it can kill it. Use aida64 and test again.

[Sneakyfox169]

A little off topic, but where did you decide to put the res/pump in the case? Pics?

OT: Don't use prime95 on your CPU; it can kill it. Use aida64 and test again.

 

I'm using a pump/res combo that slots in the middle bay, (yeah it's a bay res). Tried aida64 stability test and the temps gave the same results. Maybe 5 degrees or so cooler, but still ridiculous temps for a cooling loop. I've also made a build log that's go pics and I talked about the steps I took when building it : http://linustechtips.com/main/topic/402355-wolfs-pride/

[Sneakyfox169]

So I've just put a Cooler Master Seidon 120M for the moment on the 5820k until my replacement pump gets here, so far at stock speeds at idle I get max temps of 42 42 41 41 42 41 considering it's currently 35 degrees ambient, I'd say it's pretty normal with AIDA 64 under stress test after 17 minutes with internet browsing I get max temps of 55 51 56 54 55 52, I'm gonna try to OC it now and see what gives.

[Sneakyfox169]

Ok OC'd at 4.5 Ghz @ 1.292v, I had it running at 1.286 the other day but I had some issue there as I tried booting it with any OC, I had to switch bios, then switch back and then it booted while oc'ed, not sure what was going on but I raised the voltage a tiny bit to give me a bit of headroom for stability. For idle with corsair link, cpu-z, realtemp gt and internet open, youtube vid playing and me on this forum I have max temps of 48 46 50 50 46 44, since turbo boost is disabled it's running at full power constantly for a single 120mm AIO I think that's pretty much normal no ? Gonna try aida 64 now, I'm hoping for max mid 70 at tops, if it goes to 78 or more I'll stop it. Let's see how it goes... Ok... Well after about 30 seconds I got 79 73 87 75 82 71 degrees for the core temps, though it was all over the place, went to 80+ then back down to 68 then back up, no consistency, sooo not sure if normal or not... your thoughts are welcome and needed...

[TDP_Equinox]

Ok OC'd at 4.5 Ghz @ 1.292v, I had it running at 1.286 the other day but I had some issue there as I tried booting it with any OC, I had to switch bios, then switch back and then it booted while oc'ed, not sure what was going on but I raised the voltage a tiny bit to give me a bit of headroom for stability. For idle with corsair link, cpu-z, realtemp gt and internet open, youtube vid playing and me on this forum I have max temps of 48 46 50 50 46 44, since turbo boost is disabled it's running at full power constantly for a single 120mm AIO I think that's pretty much normal no ? Gonna try aida 64 now, I'm hoping for max mid 70 at tops, if it goes to 78 or more I'll stop it. Let's see how it goes... Ok... Well after about 30 seconds I got 79 73 87 75 82 71 degrees for the core temps, though it was all over the place, went to 80+ then back down to 68 then back up, no consistency, sooo not sure if normal or not... your thoughts are welcome and needed...

It jumping up and down is normal for intel CPUs, it's different for AMD CPUs.

As for those temps, they seem fine.

[Sneakyfox169]

Alright, well then at least that's some good news, though it confirms that something is badly wrong with the loop, since a 360 and a 280 thick make no difference temperature wise, I'm probably gonna have to wait around two weeks at best before I get a new pump, so I settled for a 4.0 Ghz OC @ 1.23 v for the time being, I just tried random values it's stable so I might be able to lower the voltage a bit but I'm not too interested in spending too long trying to get an OC for a temporary solution, we'll see. Well until I get my new pump there's not much more to do I guess... Final temps report @ 4 Ghz are under Aida 64 test 75 73 79 75 78 70.

[TDP_Equinox]

Alright, well then at least that's some good news, though it confirms that something is badly wrong with the loop, since a 360 and a 280 thick make no difference temperature wise, I'm probably gonna have to wait around two weeks at best before I get a new pump, so I settled for a 4.0 Ghz OC @ 1.23 v for the time being, I just tried random values it's stable so I might be able to lower the voltage a bit but I'm not too interested in spending too long trying to get an OC for a temporary solution, we'll see. Well until I get my new pump there's not much more to do I guess... Final temps report @ 4 Ghz are under Aida 64 test 75 73 79 75 78 70.

Those temps are fine, but for a temporary solution I would remove the OC all together.

[Sneakyfox169]

Well I've just received a package from the cooling retailer I bought the pump from, it looks new, unfortunately I'm not home yet so I can't test it yet, but it was faster than I was expecting. I say it looks new because I didn't get a reply from the retailer. They received my package and on the same day they sent something back. If it's a new pump, well I'll be bloody impressed. That said they didn't send any e-mail saying the pump was found defective or anything, not my first time with customer support, but the first time I get something shipped back the same day with no info. So yeah I guess that' pretty much it. If putting the pump in doesn't fix my temps I'll try running the loop backwars so I can put the cpu block on the right side up. If that still doesn't work I'll have to contact EK. I guess that's it for now.

[TDP_Equinox]

Well I've just received a package from the cooling retailer I bought the pump from, it looks new, unfortunately I'm not home yet so I can't test it yet, but it was faster than I was expecting. I say it looks new because I didn't get a reply from the retailer. They received my package and on the same day they sent something back. If it's a new pump, well I'll be bloody impressed. That said they didn't send any e-mail saying the pump was found defective or anything, not my first time with customer support, but the first time I get something shipped back the same day with no info. So yeah I guess that' pretty much it. If putting the pump in doesn't fix my temps I'll try running the loop backwars so I can put the cpu block on the right side up. If that still doesn't work I'll have to contact EK. I guess that's it for now.

Odd.  Would be interesting if it is a new pump..