SOLVED EK R9 290x Water Block Problem

[0112358]

PROBLEM SOLVED: disconnected PWM cables, and the pump works properly now.

Should've thought of that.


Hello,

I've contacted EK but they haven't gotten back to me, so I'm posting here in the hope that someone has an idea about how to fix this.

I have a loop with just a pump, two full-cover EK 290x waterblocks, and a 60mm 280 radiator.

I'm using dual serial 3.2 PWM pumps, and the radiator size is 324x144x60mm.

There's only one way I can run the tubing through the GPU blocks that allows the pump to work - when hooked up this way, however, the cards aren't cooled - 60º idle, and then straight through the roof until they die when I try to do anything interesting with them. I don't believe that the water is properly flowing through the blocks, since the cards are hot to the touch, but the water in the tubing directly above is lukewarm at the hottest. I also can't see how it's a configuration that should work, even though I thought that any port could be used as an inlet or outlet. Otherwise I would think I had bad contact between the GPU and the block, but I don't think that's it.

I've attached a picture of what is and isn't working. Any ideas?

 

[skorpien]

You want either the second or third loop setup. Never the first or fourth, as in those situations, at least one block will NOT be cooled. (The second pic is serial, the third parallel.)

That said, it would seem your pump is insufficient to power your loop. What pump are you using?

Also, how much radiator are you using? You need to provide more information in order for us to make the best recommendations.

Edit: To elaborate on my first point, the water in the loop will take the path of least resistance. In the first case, it will bypass both cards (lowest restriction, but least effective obviously. This would explain your sky rocketing temps as neither card is being cooled). In the fourth case, the bottom card is being cooled, but the top card is being bypassed.

Now that I think of it, I'm sure it's your pump, as it seems to have no problems with the least restrictive setup but chokes whenever restriction is introduced.

Edit 2: I highly recommend you read the stickies in this subforum. They're a great introduction to watercooling and will help you gain a better understanding of the hobby.

 

[0112358]

Thanks for your reply - I've added additional info above.

 

[skorpien]

That is really strange. Those pumps should have no problem with that loop. Are you sure the loop is being bled correctly (getting as much air out as possible)? Is one of the pumps dead?

 

[GTXJackBauer]

Welcome to OCFs!

After reading this thread my guess is that your pumps aren't pushing any fluid.

Here are a few questions to start with:

Are you using 2 pumps in that reservoir or 1 pump?

If you are using 2 pumps, are the pumps in serial or parallel?

What speed are the pump(s) set at?

Do you hear loud noises or a low humming sound from the pump(s)?

Do you see the liquid moving in the reservoir?

How old is this loop and did the reservoir accumulate any bubbles when bleeding the loop?

From my past reads, those reservoir seem to be a pain in the wazoo to prime the pumps and bleed the loop. I would look more into that area as the culprit. Your temps are too high and if your tube feels a bit warm than my guess is that there isn't any flow or barely any for that matter. Try touching the tubing furthest away from the heat source (GPUs).

 

[Tatsugen]

Aw man, I just put those blocks on my GPUs today... And I'm going to be bleeding it tomorrow...

Just a side note too, I say on EK's website, they were saying not to use more than 5mm thread fittings on the blocks. I don't know of many that are over 5mm, but maybe that is causing a block?

 

[0112358]

They're good blocks - nothing wrong with them, don't worry.

 

[EarthDog]

Outside of the issue in the OP, has anyone else noticed that it is only a 2x140mm radiator to cool 600W of heat???!

EDIT: Why is this marked as solved? I do not see a resolution. If there is a resolution, please post it up so those that run across the thread can see what was done.

 

[Tatsugen]

Outside of the issue in the OP, has anyone else noticed that it is only a 2x140mm radiator to cool 600W of heat???!

Extremely high rpm fans? Even so, the cards shouldn't hit 100C... I should find the post (look below, found it), but someone has done a dual 180mm, and you get high 60's low 70's C at load... with low rpm.

Edit: Heh, somehow pressed the post button before finishing typing post...
Edit 2: Found the link, actually a fan review, but he only has a 180. in the case.

 

[EarthDog]

You are missing my point...I am aware that is/was not the cause of the issue. Just saying that, to get better than air temperatures with a 'quiet' environment, that is not nearly enough rad.

Seems pointless to me to watercool to get air like temps.

 

[Tatsugen]

Ah, sorry for the misunderstanding. That's why I was getting convinced to put more radiator space in... Though, anything below 80C was better than my stock air cooled cards.

So, recommendation, to get more radiator space?

 

[EarthDog]

Its probably best to discuss this in your already existing thread...

 

[moho]

is it me or is all of the diagrams not making any flow through the blocks?
shouldnt it be:

 

[EarthDog]

The whole thing seems backwards to me... What I mean is he has the pump feeding the BOTTOM card on up.

I would have the pump out feed the top card inlet. then out to the next card, then out to the rad...

 

[Tatsugen]

@EarthDog
Yeah, I guess he really wasn't asking about temps... just flow.. And I think I have enough rad. But back to 0112358's topic...

@moho
I think he was just testing to see if he could get any flow at all. So, yes, there wouldn't be any flow through the blocks, but you would get a less obstructed loop flow.
The two middle diagrams in his original post are ones that should be good flows.

 

[skorpien]

is it me or is all of the diagrams not making any flow through the blocks?
shouldnt it be:

That's technically the same as the second diagram, the only difference is that you're reversing the flow through the individual blocks.

 

[madcratebuilder]

The whole thing seems backwards to me... What I mean is he has the pump feeding the BOTTOM card on up.

I would have the pump out feed the top card inlet. then out to the next card, then out to the rad...

No kidding, In SLI/CF the primary card, or top card runs hotter from my limited experience.

 

[GTXJackBauer]

No kidding, In SLI/CF the primary card, or top card runs hotter from my limited experience.

That all depends on the configuration set on the loop. Parallel will get you even temps on both as series will get you a warmer top card.

 

[EarthDog]

Considering the temps within a loop rarely vary outside of 1-2C, is the difference really that big?

Has anyone figured out how this was solved?

 

[GTXJackBauer]

Considering the temps within a loop rarely vary outside of 1-2C, is the difference really that big?

Has anyone figured out how this was solved?

PROBLEM SOLVED: disconnected PWM cables, and the pump works properly now.