Water pump pressure drop help OR when should the water loop be closed.

[ddfault]

When I turn on my Swiftech MCP35x water pump with it open to the atmosphere I notice the water level in the reservoir above it drops. When the pump is turned off the water level in the reservoir rises. I do understand water block will increase the resistance to flow of the water based on their specific design and the water flow rate.

Should I fill the reservoir with the pump on and then seal it or fill the reservoir with the pump off and then seal it.
Is it advisable to leave the system open to the atmosphere so it will normalize pressure in both situations?
Which pressure level is the recommended one?

Here is my setup. The water goes from the reservoir to pump to radiator to chipset to cpu to gpu and then back to the reservoir.



Heres a short video example of what I'm talking about. The clear tube is open to the atmosphere.
The MCP35x starts at a low flow rate, it increases to max flow rate, and then returns to a low flow rate. In the middle of the video you will see the water in the tube drop and towards the end it will rise back up.



Any advice would be helpful.
Cheers

[Waterlogged]

That is drop due to lots of air trapped in the rad. Your going to have to tip the rad around bit to get it out. Had you run the pump to the lower port on the rad, you likely wouldn't have had this problem.

[CedricFP]

Yup. Water is not compressible, but air is (IIRC). So when you turn your pump on, the air compresses and you experience a water level drop.

The worst thing about it all is that you're not utilising your radiators to their fullest capacity.

[ddfault]

Thanks for the quick replies! I did actually know water doesn't compress, but for whatever my mind didn't link the two. Thanks for pointing out the obvious there.
I have a port on the right side of my radiator...



So I unscrewed it from the top of my PC and with that side up opened that port. If there was air on that side, there wasn't a huge amount. I put cap back in with water leaking out in an attempt not to add more air.
I turned the radiator up the other way to see if any air bubbles made their way into the tubing, but none showed up.

With everything put it back together, the water in the clear tube still sucked down into the reservoir when the pump was turned back on. I have since removed the clear tube and its fitting from the top of the reservoir, filled the reservoir completely full while the pump was running, and put the original fill cap back in place in the reservoir top's opening. During the 24 hour leak test water bubble did keep showing up on the top of the reservoir. I would open it, add water, and later some more would appear. I thought one of my o-rings might have been loose, but nothing ever leaked enough water for me to see.

Will time help all the air get out of the system?
In this state should I fill it with water while running and the air is compressed or when the pump is off and the air is not compressed.

I'm using a little over 1 quart of distilled water with around 5 cap-fuls (teaspoons) of water wetter.

[Waterlogged]

Water wetter?...Flush the system now! That stuff is crap and hasn't been used by anyone that knows better for quite a number of years.
While you've got the loop drained, swap the tubes on the rad and that should take care of your problem.

[ddfault]

Ok ok I'll flush the stuff. What do you mean swap the tubes on the radiator? I'm using EK nickel plated water blocks.... so what should I be using now that my silver coil isn't recommended.

[Waterlogged]

Put the tube that's on the top port on the bottom and vice versa.

Something other than EK Nickel Blocks. That whole report that they had done stunk worse than skunk spray by page 7, they spun the data so that it was favorable to their standpoint that it was the end users fault. Fact of the matter is, there are other companies that have nickel plated blocks out there and none of them have the problem EK has.

[ddfault]

Ok, I will swap tube connections on the radiator. I will drain, clean, and refill the loop tomorrow. So perhaps in another day or so we can revisit the situation.

I was meaning what additive should I use instead of the Water Wetter with the distilled water? I've got 23 months to kill the EK water blocks .

After the water block the water hits my reservoir which has the sponge and stainless steel mesh. If I get flaking, I'm thinking I'll notice it in the sponge.

I read the EK official statement, halt on EK nickel plated blocks, and EK RMA upgrade for corrosion. It sounded to me like they are trying to do right by the customers.

[CedricFP]

I don't use any additives with my loop. Just silver plated barbs.

And yes, having the radiator's outlet be the highest port is the best idea.

[ddfault]

I usually use dish soap like dawn, hot water, and a toothbrush to clean parts. Is this the suggested tools?

[CedricFP]

I tend to just flush with hot water (rads) or use ketchup to clean oxidised copper.

Otherwise, no washing materials. How dirty can your parts be in a closed loop? If you're worried about that water wetter, and you want to be super anal, flush your components in hot distilled a few times over.

What on earth are you cleaning with a toothbrush too? Like, gunk in your block or..?

[ddfault]

When I was reusing an old pump from a friend's attempt at water cooling I noticed that parts got coated in a white film when they dried. Cleaned them with the hot water, soap and toothbrush then to get it gone.

Gotta clean those tiny crevices somehow

EK Supreme High Flow CPU

EK-VGA Supreme High Flow


I was assuming I would do the same to get whatever film the water wetter might leave behind.

If you think flushing distilled through 2 or three times will do the trick I'm all for not re applying thermal paste and leak testing YET AGAIN.

Wouldn't plain white vinegar mixed with water work better than ketchup? No sodium, tomato paste, or high fructose corn syrup in the vinegar