Really dire temps for new loop - advice...

Just ran Heaven for 10 mins and CPU went up to 57. Card 1 went up to 79, card 2, 68.

Pipes were warm. Backs of the cards felt hot.

As far as I know, the raystorm block has no in and out. Correct me if I'm wrong.

I didn't clean the stuff off the actual gpu when the cap came off, I just put some MX4 on top of it. It was dark grey, looked nothing like thermal paste, I didn't know what it was.

How deep are the cards? I don't know what you mean.

When I turn the pump up or down, I hear no difference whatsoever in sound. Flow doesn't seem any different. Pump doesn't get any louder or quieter, and the red screw on the back doesn't feel right at all.
 
I've decided to contact EK for a block RMA. I agree with ttl, temps are ridiculous, and there must be a fault with the block. I can use this thread as evidence if they ask

Will be back on here when I've got replacement and everything is sorted, posting results.
 
It's easy to remove the back cover from a D5 with a flat blade screw driver in order to check. In my last project I desoldered the 10k pot that the red speed screw goes in to and soldered back in the wires for a new 10k pot that I placed in a more convenient location. As I had 2 D5's I used a double gang pot so that the speed of both pumps changed by the same amount when the external knob was turned.

Hope you get it sorted mate, it's frustrating when your work doesn't pay the dividends you were expecting.
 
Honestly I don't think its the blocks, after all they are only pieces of copper.. What could be wrong.

If your tubes and blocks are warm that can only really mean two things: You are flow limited or you do not have enough rads. Water in your loops should flow so fast through your blocks that it barely gets warm. Considering the rads you have your temps should be lower. So i think your flow is really limited somehow.

Like SuB said, make sure the threads on you barbs do not go deep enough into the block to cover the ports leading in to the GPU block itself. If that made any sense. This could also the the reason you don't see any difference in flow.
Or that your pump is stuck on low somehow like stated before..

oh, and how fast does your cards get up to those sort of temps?
 
Last edited:
The reason I think it's the block is both cards were working fine in my rig up until about a week ago on air. It was always the top card getting very hot, and it just so happens that is the card where the stand off had fell off and the thread had stuck in the block itself.

I can't go back to air, because in the mountain of gear that's been floating around in my office this past week during this balls up, I've lost the screws for the air cooler.

This whole project has been a complete disaster from start to finish. It's put me completely off water cooling. I wish I'd stayed on air and just gamed with headphones on.

Sorry to throw the proverbial toys out of the pram lads but right now I just feel like ****ing the rig off out of the window and buying a laptop.
 
100% mate blocks dont cause that much difference.

If youve lost the cooler screws then youre screwed tbh.

You can rma cards you just refit the old cooler and hope for the best.
 
If the standoff came away from the block and it wasn't refitted properly it could cause poor contact. When I first put my blocks on my cards I killed the first one (no idea, worked fine after re-seat) and the second I hadn't screwed a standoff in enough after it fell out, causing the ram to overheat/ spaz out.
 
Good to see that your on to something, defo something wrong with the pump. Is it the pump making that noise?? doesn't sound good at all...
 
Yes. Sounds like my missis vibrator.... apparently.... :/

20130126202133.jpg


That's broken stand off. I'm hoping once I get the pump RMA'd that its not going to cause a problem again. Only time will tell. Will report back once I have the new pump and a response off EK

Time for a beer. Well, well earned one!
 
Since that pump has a rpm wire it would be interesting what it was reading. Have you taken the back off the pump to check the pot or looked at the small ceramic bearing? The impeller thingy could be jammed? It would be a Dynamo magic trick if you looked in the pump and found your blocks stand-off :)

You might not have to send the pump back to Anne Summers ;)

Jusy watched your vid with sound... that pump sounds like it's running dry and at max speed. Is there something blocking inside the res?
 
Last edited by a moderator:
I get lower temps with air cooled 480's, that is really bad for WC... no offense meant.


Do remember, you are cooling Fermi, not GCN or Kepler, the heat output of these GPU's is large, although, i have no experience in watercooling, so i shall shut up and escape further embarrassment.
 
Well, I officially gave up trying to diagnose the problem today and f****d the cards off in the cupboard till I can be bothered to RMA them. The bottom one just gave up the ghost on it's own (completely stopped working) and the top one continued to show ridiculous temps for a w/c loop, so I took it out.

With the money I have left from my PPI claim, I've bought an EVGA 690 and XSPC waterblock. Finally - a card that will actually work. I've forgotten what it feels like.
 
It was for the best anyway mate.

You probably have an airlock in the pump,try a single drop of dishwashing liquid in the res.
 
Well, it's DONE!!!!!!!!!!!!!

Got the card today, installed it, finished the build. Temps under load, AT LAST!!!!!

GPU 1: 35 C
GPU 2: 37C

Immense. And I am so happy that after all the effort it finally works. Eureka!!!!!
 
Hi,

I know I'm rather late to this thread! Firstly, well done getting to the bottom of the problem. As I was reading I was thinking it must be some flow issue, likely pump...it was nice to see you got there too! :) I actually tested my loop while flushing it out, so clean deionised in > loop > bucket so I could see that the flow was excellent.

I would like to comment however on your initial setup using parallel flow on your GPU's. I'd not expect you to get the best temps out of that, as you had it configured a little wrong. To get parallel flow to work at its best you need the in and the out to both be in the top or in the bottom ideally - just my observation from much playing/experiementing.

In my own rig, the coolant comes in the TOP into the first card and through the second, before exiting once more via the top card and on to my CPU. Obviously as it's parallel flow, roughy half the collant that goes in the top card exits out and up, whereas the half that enters the lower card (not getting heated by said card of course) has a near straight line exit back out again.This setup works very well.

I will say however - indeed I said it in another thread - those metal adjustable SLI connectors are really quite restrictive. I've compared my Phobya SLI connectors vs. some smaller flexible pipe and the latter wins. The adjustable connectors make a far neater build for sure, however my first GPU might hit 38c during extreme benching, the 2nd GPU will be 4-5c higher than that consitantly. Compare this to the smaller hose (I use 1/2" ID throughout my system except for these smaller bits of hose) and I see 38c of BOTH GPU's under the same conditions.

As the flow is parallel overall flow is just fine, water will find the easiest way, and the loops does work just fine. However my imbalanced temps do bug me a little.

Note: serial flow works just fine of course and your temps are particularly close anyway so that's great. Personally, for my first WC build, I was keen to play around with parallel flow as I thought it quite an elegant solution. It's just a shame those SLI links cannot be had in a larger diameter - ideally 1/2" like my tubing.

I will add that my loop in near silent, as I'm very much with Tom on that side of things. Likely I could improve my cooling with the addition of a bit more noise, however, considering I rarely go more than 20c over ambient on the GPU's and 30c on the CPU, well, I'm fairly content in that regard. Oh, I thought my LAPTOP had a problem as the thing seems quite noisy now - silly I know, but true none the less!

Anyway, I think you've had your baptism of fire..erm, water now, so can enjoy your rig with the satisfaction of a job well done.

Cheers,

Scoob.
 
Back
Top