GPU Water Block Delaminating

[VegArchie]

Hello,

I purchased the Eisblock Aurora GPX-N GPU Waterblock for RTX 3090/3080 (Plexi with Backplate) in December. I installed into an existing system that had a CPU block, pump, reservoir and radiator. The system had been working fine for two years. All components are copper based (not aluminum).

A few days ago, the CPU and GPU temperatures started to rise and I noticed that the flow of water had slowed down considerably. After rebooting the computer, there was no flow at all. I inspected the pump and found that it works fine. After draining the system, I was able to determine that the input on the CPU block (the first point after the GPU) had completely clogged with debris (green spikes or crystals that clumped together). I was able to locate the origin of the green spikes/crystals to the middle of the GPU water block. To me it seems that the nickel plating, in the GPU block, is coming off and the copper is corroding and the plating is what clogged and destroyed the system. I can find chunks of this in the pump and in the radiator as well. The system is shot. I have had to switch to air cooling. I have attached a picture of the crystal clog in the GPU block as well as the CPU block showing the clogged input and clear output.

Can someone please get back to me to resolve this issue?

 

[Eddy]

The nickle plated part looks good, but there is a lot of dirt inside. If the nickel plating comes off, it looks different. But there is something in your circuit that causes a lot of dirt. Either the wrong fluid or the fluid has been tipped. Or there was a massive amount of dirt in the system somewhere. You have no choice but to clean the whole system properly.

 

[VegArchie]

Perhaps it's not delaminating, but I don't understand how it could be dirt. I find this explanation to be lacking evidence. As I mentioned, the system existed and ran perfectly for two years. I performed regular flushes every 3 months. After I installed the GPU water block, I flushed the system with a cleaner. Then filled and used the same brand water coolant was I have always used. Where would dirt come from? What kind of dirt? Whatever it is, it's emanating from the GPU block. It can't be cleaned as it's completely clogged.

Is it corrosion? It very much seems like a malfunction of the GPU block. Please help.

 

[VegArchie]

What does "fluid has been tipped" mean?

 

[Eddy]

With some things it is difficult to translate something from German into English, sorry. I meant that the pH value of the liquid must have changed massively or something got into the water that turned the liquid into an acid that attacked the nickel plating. This can happen through all sorts of things.
Corrosion can only occur if you still have aluminium in the circuit or other metals apart from brass, copper or nickel-plated copper. What detergent did you use for flushing? Because you should actually do without detergents. The residues can also cause this.

 

[VegArchie]

I don't know of any instance in which a closed loop system can have something get into the water.

I did not use a detergent. I used Primochill Sys Prep. It is the same fluid I used when I setup the system. It did not cause problems in the past. The manufacturer also states that "Residual treated water left in your loop is not harmful." So I do not believe that this is the problem.

If the pH changed, there would be corrosion everywhere. However, the corrosion is happening on the plate of the GPU block and then breaking off and clogging the "entry" to the CPU block, as shown in the pictures. The "exit" of the CPU block is clear, with no corrosion, indicating that the CPU block was performing normally with the water coolant.

In all honestly, the pictures look exactly like what would happen if the GPU block was aluminum. Is there any chance, your manufacturer messed up and made the block with aluminum? The GPU block is the only part that changed in my system. All of the corrosion is happening on the plate of the GPU block and flowing from there downstream. I am not meaning to be accusatory, but all the evidence points to the GPU block, and now I am out $1200 in parts. There are no other parts of the system with aluminum nor has there ever been. Again, if there had been, I would have seen the corrosion before installing the GPU block.

Do you have any other ideas on what could cause this that is consistent with the evidence?

 

[davido_labido]

Any dirt will always get stuck in the GPU and CPU blocks due to those being the points of most restriction. There is very little space between the fins, so naturally the dirt clogs there, usually at the inputs of each as it is stopped from reaching the output.

There is no chance it can be aluminium, we don't have aluminium blocks.

It looks somewhat like plasticiser maybe, or as Eddy said, contaminent. It is very easy to introduce dirt to a loop, even just when you're doing system maintinence.

Is there any visible copper on any of the blocks? If the nickel is flaking off, then copper would be visible - also, nickel flaking *usually* wouldn't cause harm, it would usually just not look pretty.

As the particles are on the inlet to the waterblock, I would suggest checking the parts before it and making sure there is nothing there. A filter before the block would show if there is anything causing issues prior to it.

 

[VegArchie]

What is plasticiser? What type of contaminent would cause this?

 

[davido_labido]

placisciser/plasticsizer is used in some tubing (particuarly PVC) to make it more maluable overtime, this can break down and then can flow throught he loop. It will usually get stuck in the fins of waterblocks - very similar to what you see here. There is a lot of information about this online if you wish to research it further.

It is always difficult to track contamination down as it can be caused by so many things, if the PH levels change, it can cause nasties too. It really is one of those things that is very hard for us to advise about as there are usually many different brands, cleaning agents, coolants and environments that change the values and we can only really help with the products and combinations we have tested.

We will try our best to help you though as many of us started as liquid cooling enthusiasts so we have certainly seen our fair share of issues!

 

[VegArchie]

Looking at the tubing that was added, (PrimoFlex Advanced LRT Soft Flexible Tubing -3/8in.ID x 1/2in.OD, 10 feet - Crystal Clear), it appears to be intact, and the pictures I find on the net regarding placisciser issues in a loop look different form what I see. The gunk from the plastic is a green slime/goo clog, whereas the gunk in my loop is a faint, green crystal formation. Similar to the color of copper when it is exposed to air/water.

Is it possible to clean the insides of these blocks?

 

[davido_labido]

Yes, you can take the blocks apart to clean them. You can also flush fluid the oposite way which usually will dislodge the gunk. As far as I know, LRT Advanced is plasticisor free, but I can't be 100% as I haven't used Primochill in around 7 years!

Green is usually a sign on copper reacting with something, it could be a number of things however. A good clean/flush through and then some new fluid would be my suggestion.