Concept Ghost S1 DDC Top + Reservoir + Distro Combo Idea

[jay8ee]

** Update 12/05/2020 **

I have decided to remove the GPU distro pipes from the design, this is for a couple of reasons:

1. Actually mounting the plate to the GPU is very challenging, and getting this right is probably going to take a lot of time and effort.
2. This would have only ever work for one very specific GPU block.
3. Never sat right with me that the GPU pipes were reversed, now you can choose which way it is hooked up.

** End of update **

Hi all.

I've been tinkering with this idea of a custom water block for the S1. From the outset my design goals/constraints have been:

• Include a DDC pump
• Include a reservoir to keep pump primed and make loop prep easier
• Rad pipes must be at the front (because they look cool IMO)

I've not given much thought to the screw placement yet, I know that at present there probably isn't enough clamp pressure throughout. I am continuing to play around with this.

I would appreciate any feedback on this. Please note that I posted this on Reddit, too.

 

[bricemx]

Was thinking about distro plate and pump combo for long and here it is. Look very nice, it at the right place also, in GPU compartment and keep update with more image till final design. Thanks to make a project like that for ghost S1

 

[oOflyeyesOo]

Hopefully it can be flipped to have the tubes in the back IMO.

 

[jay8ee]

Hopefully it can be flipped to have the tubes in the back IMO.

The tubes are connected to the back of the distro already.

 

[flar]

Is that driving the GPU block in reverse?

 

[jay8ee]

Is that driving the GPU block in reverse?

Yes. EK says it doesn't make a difference for that block.

 

[flar]

Yes. EK says it doesn't make a difference for that block.

Well, 1.5c to be more precise. Not a big deal, though.

Another factor here is that the flow is driving the GPU first. I have a GPU->CPU flow in my system and I wish I'd done it the other way around. What I'm finding is that the GPU barely gets hot at all with watercooling but it dumps by far the most heat into the system. The CPU, on the other hand, gets really hot and it dumps very little into the system. This is based on running GPU-only or CPU-only benchmarks and watching both component temperatures and water temperature. The water temps are much higher (just under 50c) when running the GPU benchmarks. The GPU temps remain under 50c, though with occasional forays into the lower 50s. With CPU benchmarks the 3900x can near its thermal limits, but the water temp barely gets to 40c. It may not make a huge difference, it would depend on the temperature drop through the radiator, but I would rather have the component that is most in danger of hitting its thermal limits get a shot at the fresh water than the component that never seems to get anywhere near its limits regardless of how much heat is dumped into the system.

The CPU results could also be due to a bad seating of the block, but I've already pulled apart the system twice while setting it up and the thermal compound was pretty well spread both times I did an inspection, and the thermal results were always consistent.

 

[jay8ee]

Well, 1.5c to be more precise. Not a big deal, though.

Another factor here is that the flow is driving the GPU first. I have a GPU->CPU flow in my system and I wish I'd done it the other way around. What I'm finding is that the GPU barely gets hot at all with watercooling but it dumps by far the most heat into the system. The CPU, on the other hand, gets really hot and it dumps very little into the system. This is based on running GPU-only or CPU-only benchmarks and watching both component temperatures and water temperature. The water temps are much higher (just under 50c) when running the GPU benchmarks. The GPU temps remain under 50c, though with occasional forays into the lower 50s. With CPU benchmarks the 3900x can near its thermal limits, but the water temp barely gets to 40c. It may not make a huge difference, it would depend on the temperature drop through the radiator, but I would rather have the component that is most in danger of hitting its thermal limits get a shot at the fresh water than the component that never seems to get anywhere near its limits regardless of how much heat is dumped into the system.

The CPU results could also be due to a bad seating of the block, but I've already pulled apart the system twice while setting it up and the thermal compound was pretty well spread both times I did an inspection, and the thermal results were always consistent.

Interesting. In my research, I generally found that loop order didn't seem to matter, but I am no expert by any stretch.

When I get some time I will test my build both ways and see if it makes a difference.

 

[flar]

Also, those considerations may vary depending on GPU and CPU. I'm using 2080Ti on Ryzen 3900x. Intel processors tend to run at higher overall wattages when pushed, don't they?

I also need to see if the CPU throttles more or runs hotter based on the water temp. It may care more about total wattage leeched than the specifics of the temperature at which it happens. My GPU is currently not in my loop as it was just returned from an RMA so I will try to run some tests before I reinstall the GPU block. In the meantime, I'm letting the replacement get some Folding done to make sure I won't have to do another RMA...

 

[Patongo]

Well, 1.5c to be more precise. Not a big deal, though.

Another factor here is that the flow is driving the GPU first. I have a GPU->CPU flow in my system and I wish I'd done it the other way around. What I'm finding is that the GPU barely gets hot at all with watercooling but it dumps by far the most heat into the system. The CPU, on the other hand, gets really hot and it dumps very little into the system. This is based on running GPU-only or CPU-only benchmarks and watching both component temperatures and water temperature. The water temps are much higher (just under 50c) when running the GPU benchmarks. The GPU temps remain under 50c, though with occasional forays into the lower 50s. With CPU benchmarks the 3900x can near its thermal limits, but the water temp barely gets to 40c. It may not make a huge difference, it would depend on the temperature drop through the radiator, but I would rather have the component that is most in danger of hitting its thermal limits get a shot at the fresh water than the component that never seems to get anywhere near its limits regardless of how much heat is dumped into the system.

The CPU results could also be due to a bad seating of the block, but I've already pulled apart the system twice while setting it up and the thermal compound was pretty well spread both times I did an inspection, and the thermal results were always consistent.

I'd like to clarify that at low volume of water it probably make a difference in relatively low ambient temp, but delta T across wattage is subjected to parameters such as water volume, flow rate, fan speed, ambient air, block thickness, fin density, etc.

 

[jay8ee]

Hmm, this is now a more general discussion about water cooling it seems. Appreciate points and feedback, I'm going to be continuing with this project and will report back with my progress and findings as I have them.

 

[jay8ee]

Decided to rework the design a bit. Please see the original post for the update.

 

[flar]

I'm looking for a good reservoir to eventually use with a redesign of my water loop if I can ever get my hands on a 3080/3080 Ti FE with the short card. It looks like this reservoir/distro will work even with a longer GPU, but it looks nicer and would fit better than anything else that I can find that shoehorns into the end of the GPU space.

Is this still under development?