New and Improved Watercooling Sticky - Post Your Systems Here

I need deets on that GPU+Water block. Full active waterblocks on both sides!?
Yes its an EK quantum vector for 3090 TUF https://www.ekwb.com/shop/ek-quantum-vector-tuf-rtx-3080-3090-d-rgb-nickel-plexi
And it has the active backplate: https://www.ekwb.com/shop/ek-quantum-vector-re-rtx-3080-3090-active-backplate-d-rgb-plexi

I think they still make them even for the 4090 but its probably more relevant / needed for the 3090 due to the ram being on both sides of the card.
 
Last edited:
Yes its an EK quantum vector for 3090 TUF https://www.ekwb.com/shop/ek-quantum-vector-tuf-rtx-3080-3090-d-rgb-nickel-plexi
And it has the active backplate: https://www.ekwb.com/shop/ek-quantum-vector-re-rtx-3080-3090-active-backplate-d-rgb-plexi

I think they still make them even for the 4090 but its probably more relevant / needed for the 3090 due to the ram being on both sides of the card.
It doesn't look like EK have an active backplate for the 4090. The waterblock I got came with a passive backplate.
20240302_095549.jpg
 
The 4090 doesn't need an active backplate imo. Just a waste of money. 3090 definitely did due to the hot G6X modules on the rear due to the clamshell design.
 
The 4090 doesn't need an active backplate imo. Just a waste of money. 3090 definitely did due to the hot G6X modules on the rear due to the clamshell design.

There is definitely some C°s to be carried away by sinking the back of the board behind the memory modules, GPU, and VRMs.
 
There is definitely some C°s to be carried away by sinking the back of the board behind the memory modules, GPU, and VRMs.
Sure...but I don't think enough to warrant an active backplate. Just my opinion. Considering how expensive EK stuff is, seems a waste to me.
 
Sure...but I don't think enough to warrant an active backplate. Just my opinion. Considering how expensive EK stuff is, seems a waste to me.

Well, there isn't really a point in spending money in half-assing the project when you're already spending a grand to block the box
 
My MiL asked 'why watercool' after the latest build. I thought a second and admitted it's really just for fun these days. It's gotten much easier, but also less needed.
 
My MiL asked 'why watercool' after the latest build. I thought a second and admitted it's really just for fun these days. It's gotten much easier, but also less needed.
Very low noise with good temps isn't possible on air if you have a 300w or higher GPU releasing heat inside the case. You need to spin up case fans and CPU cooler fans to compensate when on air while water keeps everything cool at low noise as long as you have enough capacity. If you are running a 200w GPU then it is fairly easy to handle, but you still need a beefy tower to not have to spin up the CPU cooler fans quite a lot on modern CPUs and you won't get below moderate noise. My main system tops out at around 31dba at 20cm from the case in the summer and stays below 30dba in the winter, which would be impossible on air as the components would be boiling.
 
Yeah, Less needed, not unneeded. My main system Folds any time I'm not using it, so it might as well be benchmarking. Still pretty quiet, and really quiet if I pause Folding@home...
 
It pretty much is unneeded at this point. I dumped my loop when I updated to a 4080 and 7800x3d with all air cooling. Yes temps are a little higher but my PC is still pretty silent still.
 
It pretty much is unneeded at this point. I dumped my loop when I updated to a 4080 and 7800x3d with all air cooling. Yes temps are a little higher but my PC is still pretty silent still.
Define pretty silent. Mine isn't silent, just low noise at around 30dBA under full load. Pretty silent to me is you can't hear it while listening for it as it is drowned out by the noise floor in a quiet room, typically around 26-27dBA measured properly (no phone apps or similar as they are generally very inaccurate due to mic quality, lack of calibration etc.). Low noise is you barely hear it when listening for it. If you are at around 35dBA then I would consider it in the moderate to somewhat loud range.
 
Define pretty silent. Mine isn't silent, just low noise at around 30dBA under full load. Pretty silent to me is you can't hear it while listening for it as it is drowned out by the noise floor in a quiet room, typically around 26-27dBA measured properly (no phone apps or similar as they are generally very inaccurate due to mic quality, lack of calibration etc.). Low noise is you barely hear it when listening for it. If you are at around 35dBA then I would consider it in the moderate to somewhat loud range.
Silent enough I can sleep in the same room with out annoying me.
 
My MiL asked 'why watercool' after the latest build. I thought a second and admitted it's really just for fun these days. It's gotten much easier, but also less needed.
It's also an addiction. I've never wanted to go back to an air cooled PC after my first foray into custom loop in 2011/2012.

As for noise...we do have air cooled PCs in the house and yeah, mine is definitely quieter by a lot...
 
My loop was filthy so I spent the weekend cleaning it up. I built this system at the end of last year but for what ever reason my loop was contaminated after my previous attempt to clean it (which was the second attempt).

I took the opportunity to lap my CPU waterblock and changed the thermal paste on both the CPU and GPU to Thermal Grizzly Kryonaut Extreme.

To clean, I used ordinary dish soap on everything except the radiators after polishing what I could with Brasso. For the radiators I cleaned them last with a hot vinager mix (again) but this time I followed up with a mixture of caustic soda to make sure nothing survives. I then flushed with copious amounts of tap water before finally flushing with distilled water. I pre-filled the bottom radiator with coolant before installing it. It was super satisfying filling the loop this time and all the water is perfectly clear.

I also purchased a fuel filter and drilled a hole in it to fit to one of my pump's inlet so if there is ever any debris in the loop it won't clock the fins on my waterblocks.

After the rebuild I'm seeing at least 10C lower on the CPU and 1-2C lower on the GPU.
 

Attachments

  • 20240329_224244.jpg
    20240329_224244.jpg
    201.4 KB · Views: 0
  • 20240329_230931.jpg
    20240329_230931.jpg
    430.6 KB · Views: 0
  • 20240329_230941.jpg
    20240329_230941.jpg
    411.9 KB · Views: 0
  • 20240329_230955.jpg
    20240329_230955.jpg
    405.3 KB · Views: 0
  • 20240329_231129.jpg
    20240329_231129.jpg
    286.2 KB · Views: 0
  • 20240330_195234.jpg
    20240330_195234.jpg
    310.7 KB · Views: 0
  • 20240330_233239.jpg
    20240330_233239.jpg
    354.4 KB · Views: 0
  • 20240330_234522.jpg
    20240330_234522.jpg
    344.3 KB · Views: 0
  • 20240330_235645.jpg
    20240330_235645.jpg
    139.7 KB · Views: 0
  • 20240331_141054.jpg
    20240331_141054.jpg
    344.9 KB · Views: 0
  • 20240331_141102.jpg
    20240331_141102.jpg
    309 KB · Views: 0
  • 20240331_141106.jpg
    20240331_141106.jpg
    349.1 KB · Views: 0
  • 20240331_141112.jpg
    20240331_141112.jpg
    305.3 KB · Views: 0
  • 20240331_141116.jpg
    20240331_141116.jpg
    294.1 KB · Views: 0
Very nice! What are you going to settle on for coolant? And how did you rig up the filter in your res?
 
Coolant is EK Cryofuel clear. Like I said, I bought a fuel filter and driled a hole through it. I just shoved it over the inlet pipe, works great.

Just ran Timespy and got my best result. GPU temp went from (average) 56C to 49C. That is pretty incredible. CPU temps went from 71 to 68C so not as dramatic but the CPU has less load during timnespy than Prime which is similar, around 7C lower.
 
Bro that filter is fantastic. I was considering a filter but don't wanna deal with it lol. Sometimes the acrylic from the blocks breaks into fine pieces and can get caught in the micro fins. So the less I mess with the system the better cause even cleaning it I could reintroduce tiny debris again it's like chasing your own tail. A filter would introduce restriction but also keep the liquid filtered. WCing takes maintenance but it's nice when you can just let it run for 2 years straight lol. I need to run my pump at about 30 to 40 percent to get adequate flow for temps minimum. I would like to run it lower and slower but I have 4 rads and 2 blocks and a lot of angled fittings so I guess it's not too bad considering all that.
 
Bro that filter is fantastic. I was considering a filter but don't wanna deal with it lol. Sometimes the acrylic from the blocks breaks into fine pieces and can get caught in the micro fins. So the less I mess with the system the better cause even cleaning it I could reintroduce tiny debris again it's like chasing your own tail. A filter would introduce restriction but also keep the liquid filtered. WCing takes maintenance but it's nice when you can just let it run for 2 years straight lol. I need to run my pump at about 30 to 40 percent to get adequate flow for temps minimum. I would like to run it lower and slower but I have 4 rads and 2 blocks and a lot of angled fittings so I guess it's not too bad considering all that.
I don't know how much flow is restricted but there is no way it would be a 1 PSI drop. There is a pretty large surface area there. If you blow through it you can't feel any restriction at all and when I primed my pumps as usual I didn't notice any noticable drop in flow. The best thing about the placement is that the pump res drains first so I don't have to drain my whole loop to change the filter if I need to.

How are you mounting your pumps? the EK pumps I got came with two different mounting gaskets and I found there was significantly more noise with one over the other. The gasket you see in the picture (If you look closely at old vs new you can see the different gaskets) results in significantly less noise, so much so that I have one of my pumps running at 95% and can barely hear it. My pumps are running at 2200rpm (bottom) and 4700rpm (top) and they are barely audible. The other mounting gasket was only tolerable at 40%.
 
Last edited:
I don't know how much flow is restricted but there is no way it would be a 1 PSI drop. There is a pretty large surface area there. If you blow through it you can't feel any restriction at all and when I primed my pumps as usual I didn't notice any noticable drop in flow. The best thing about the placement is that the pump res drains first so I don't have to drain my whole loop to change the filter if I need to.

How are you mounting your pumps? the EK pumps I got came with two different mounting gaskets and I found there was significantly more noise with one over the other. The gasket you see in the picture (If you look closely at old vs new you can see the different gaskets) results in significantly less noise, so much so that I have one of my pumps running at 95% and can barely hear it. My pumps are running at 2200rpm (bottom) and 4700rpm (top) and they are barely audible. The other mounting gasket was only tolerable at 40%.
Yeah, the soft mounting gasket works wonders but can only be used for upright mounting. The "firm" mounting gasket lets vibrations transfer to the case. Even without the vibrations I am limited to around 55% before I can hear the pump over the fans, but with the firm one I could hear the resonances in the case from the pump at certain speeds which was intolerable.
 
I don't know how much flow is restricted but there is no way it would be a 1 PSI drop. There is a pretty large surface area there. If you blow through it you can't feel any restriction at all and when I primed my pumps as usual I didn't notice any noticable drop in flow. The best thing about the placement is that the pump res drains first so I don't have to drain my whole loop to change the filter if I need to.

How are you mounting your pumps? the EK pumps I got came with two different mounting gaskets and I found there was significantly more noise with one over the other. The gasket you see in the picture (If you look closely at old vs new you can see the different gaskets) results in significantly less noise, so much so that I have one of my pumps running at 95% and can barely hear it. My pumps are running at 2200rpm (bottom) and 4700rpm (top) and they are barely audible. The other mounting gasket was only tolerable at 40%.

I haven't gone through the trouble of any flow
meters or filters or temp sensors lol.

My EKD5 is mounted vertically with the soft rubber mount and am not worried about the noise as much as the life of the motor. I want it to last as long as possible.

I used to run it at 10-15% before with 2 rads but now with 4 rads and all the angled fittings I need to run it at least 30- 40% plus to get adequate flow for optimal temps.

The case has 24 sickle flow V2 fans that run at between 20-35% so sound is minimal and the pump is inaudible with that many fans.

I've considered adding another pump but undecided. Your build makes me want to do it because I would be able to run them slower and have more flow.

Your build is cool. I always love seeing custom dual pump builds.
 
I would be able to run them slower and have more flow.
It doesn't increase flow rate unless you have heavy flow restriction. Running dual D5's doubles head pressure but flow rate remains the same. Two blocks and three radiators isn't really enough to create a noticable drop in flow rate for a single D5. The only other scenario where dual D5's would benefit is if you have to pump water over a larger vertical distance, like running an external radiator near the floor with your PC on the desk.
 
I loaded up some timespy results if anyone is interested in the difference between stock and lapped IHS, as well as Thermal Grizzly vs Noctua NT-H1. Fan and pump speeds should be fairly consistent between runs. I haven't tried running higher fans speeds yet, they are currently between 600 and 800 rpm depending on the noise profile of the fan as I have a mix of EK 120mm and Noctua 140mm fans.

Result 1 is the current build after cleaning, lapping the CPU waterblock and using Thermal Grizzly.
Result 2 is lapping CPU only with Noctua NT-H1 thermal paste.
Result 3 is no lapping with NT-H1.
 
I loaded up some timespy results if anyone is interested in the difference between stock and lapped IHS, as well as Thermal Grizzly vs Noctua NT-H1. Fan and pump speeds should be fairly consistent between runs. I haven't tried running higher fans speeds yet, they are currently between 600 and 800 rpm depending on the noise profile of the fan as I have a mix of EK 120mm and Noctua 140mm fans.

Result 1 is the current build after cleaning, lapping the CPU waterblock and using Thermal Grizzly.
Result 2 is lapping CPU only with Noctua NT-H1 thermal paste.
Result 3 is no lapping with NT-H1.
No change with the GPU? The GPU seems to boost higher with better CPU cooling, I guess because it's getting more work from the CPU?
 
I meant "you didn't change anything," or did you lap both blocks?
Just cleaned and changed thermal paste. +250 core and +1250 memory for all runs but the clocks are boosting higher due to lower temps I guess.
 
  • Like
Reactions: Nobu
like this
It doesn't increase flow rate unless you have heavy flow restriction. Running dual D5's doubles head pressure but flow rate remains the same. Two blocks and three radiators isn't really enough to create a noticable drop in flow rate for a single D5. The only other scenario where dual D5's would benefit is if you have to pump water over a larger vertical distance, like running an external radiator near the floor with your PC on the desk.
What do you consider a noticeable drop? The drop is so dependent on the water blocks and radiators. Some radiators have flow resistance that is higher than waterblocks and there is also a large difference between waterblocks. If everything is low resistance then a D5 will be perfectly happy, but if everything is high resistance then the D5 will barely get any flow at 100% pump speed.

My CPU loop has 2 medium resistance rads and a high resistance waterblock and tops out at a little over 1gpm on full pump speed. My GPU loop has 1 medium resistance water block and 2 medium resistance radiators and can do around 1.3gpm at 100% pump speed. If I added the GPU block and one of the GPU rads to my CPU loop then I would probably top out at 0.8gpm, maybe lower. Running at 55% pump speed gives me just below 0.5gpm on my CPU loop and aprox .65gpm on my GPU loop. Thankfully my CPU block barely cares about flowrate.
 
I also purchased a fuel filter and drilled a hole in it to fit to one of my pump's inlet so if there is ever any debris in the loop it won't clock the fins on my waterblocks.
Oh nice! I used a Wix 33003 fuel filter when initially cleaning/prepping loop. Never thought to stick one in the reservoir. What model did you use?
Coolant is EK Cryofuel clear. Like I said, I bought a fuel filter and driled a hole through it. I just shoved it over the inlet pipe, works great.
Eh...careful with that stuff.
 
Why?

I've been using CryoFuel since early 2019 without issue
So when you described your loop as "filthy", what did you mean exactly?

Yeah Cryofuel once upon a time was rebranded Mayhems X1, then it wasn't and that's when it started having reports of gunking and other issues. Even the clear.
 
So when you described your loop as "filthy", what did you mean exactly?

Yeah Cryofuel once upon a time was rebranded Mayhems X1, then it wasn't and that's when it started having reports of gunking and other issues. Even the clear.

I've only ever had spotless loops :) I've had nothing but open loops since 2002 or so, and never changed coolant unless I was replacing a component.
Current coolant in my loop now is green CryoFuel that I put some clear blue UV Mayhems dye in. I think it's about 2 years old now at this point.
 
I've only ever had spotless loops :) I've had nothing but open loops since 2002 or so, and never changed coolant unless I was replacing a component.
Current coolant in my loop now is green CryoFuel that I put some clear blue UV Mayhems dye in. I think it's about 2 years old now at this point.
Part of it is I just don't trust EK QC these days. But yeah that makes sense with the colored coolant. Glad to hear otherwise no issues.
 
So when you described your loop as "filthy", what did you mean exactly?

Yeah Cryofuel once upon a time was rebranded Mayhems X1, then it wasn't and that's when it started having reports of gunking and other issues. Even the clear.
Only had issues with the clear cryofuel when my loop wasn't cleaned well enough before putting it together. Never had it on re-used components or when I've cleaned everything really well before assembling the loop.
 
What do you consider a noticeable drop? The drop is so dependent on the water blocks and radiators. Some radiators have flow resistance that is higher than waterblocks and there is also a large difference between waterblocks. If everything is low resistance then a D5 will be perfectly happy, but if everything is high resistance then the D5 will barely get any flow at 100% pump speed.

My CPU loop has 2 medium resistance rads and a high resistance waterblock and tops out at a little over 1gpm on full pump speed. My GPU loop has 1 medium resistance water block and 2 medium resistance radiators and can do around 1.3gpm at 100% pump speed. If I added the GPU block and one of the GPU rads to my CPU loop then I would probably top out at 0.8gpm, maybe lower. Running at 55% pump speed gives me just below 0.5gpm on my CPU loop and aprox .65gpm on my GPU loop. Thankfully my CPU block barely cares about flowrate.
How would I even be able to calculate the amount of resistance my radiators and blocks have without a flow meter?
I have the EK quantum kinetic D5 pump res combo and then I have the quantum vector 2 gpU block and the quantum velocity 2 CPU block with four alpha cool rads. Just looking at the EK blocks they look like they are very high resistance because of the jet plate and the tiny microfin areas. Also all of the fittings are angled and some are even double rotaries.
I don't notice a very big difference going from 30% or 40% to like for example 80% or 100%. I would say probably less than 10°?
What temperature difference are you guys seeing going from 25% to 50 to 75 to 100?
 
Back
Top