Newbie help

[fabafaba]

1 hour ago, For Science! said:

In a previous build I did I watercooled 4x 1080Ti's in the same loop, same blocks, same installation method and one card was 5 degrees hotter than the rest consistently

I think some 1080TIs just have some problem with their temp measurement. My 1080TI is at 26°C right now, but the water going into it is at 29°C, set to a 3°C delta over ambient. I checked all of my temp sensors against a thermocouple for a multimeter and a BBQ thermometer when I upgraded my loop two weeks ago, and they were all within 0.5°C. It seems like some chips/cards just aren't properly calibrated.

 

2 hours ago, Flowroro said:

that sounds valid however how are all the big youtubers who use just a single 240 rad for gpu and cpu get significantly lower temps though?

Are they using the fluid gaming stuff too? I wasn't very happy with my A240G kit and additional 360mm radiator too. I think it just comes down to the weak pump, flimsy little radiators (the "core" is just 15-16mm thick) and the aluminum blocks with a lower heat conductivity compared to copper, as well as fewer and thicker fins in the GPU block.

[Flowroro]

4 hours ago, fabafaba said:

I think some 1080TIs just have some problem with their temp measurement. My 1080TI is at 26°C right now, but the water going into it is at 29°C, set to a 3°C delta over ambient. I checked all of my temp sensors against a thermocouple for a multimeter and a BBQ thermometer when I upgraded my loop two weeks ago, and they were all within 0.5°C. It seems like some chips/cards just aren't properly calibrated.

 

Are they using the fluid gaming stuff too? I wasn't very happy with my A240G kit and additional 360mm radiator too. I think it just comes down to the weak pump, flimsy little radiators (the "core" is just 15-16mm thick) and the aluminum blocks with a lower heat conductivity compared to copper, as well as fewer and thicker fins in the GPU block.

so basically this kit is no better than air cooled and is a total waste of money? i spent about 750 aud for the whole set xd couldve just got a real copper set for a 200 more. got scammed hard, just goes to show how fake these youtubers are with their benchmarks, cant trust anyone these days

[Flowroro]

when temps are equalized, gpu is at 63c, cpu is at 50c. this is ridiculous. i can achieve better temps on AIR. something isnt right, unless i truly got scammed. I did a lot of research, everything saying its only slightly worse than copper loops. so you guys confirm I got scammed?

[fabafaba]

1 minute ago, Flowroro said:

so basically this kit is no better than air cooled and is a total waste of money?

I think it works well with an i5 or R5 and a mid tier GPU, but hardware with more oomph brings it to its limits. I spent about 350€ (570aud) on my Fluid Gaming kit, and I actually think that that is what it's worth, as long as you don't plan to pump 400W through the loop.

7 minutes ago, Flowroro said:

got scammed hard, just goes to show how fake these youtubers are with their benchmarks, cant trust anyone these days

Well, in a room with AC set to 19°C it propably works a whole lot better

[For Science!]

10 minutes ago, Flowroro said:

when temps are equalized, gpu is at 63c, cpu is at 50c. this is ridiculous. i can achieve better temps on AIR. something isnt right, unless i truly got scammed. I did a lot of research, everything saying its only slightly worse than copper loops. so you guys confirm I got scammed?

Since I cannot rule out that the block installation was somehow compromised (i.e. pic of PCB for block installation) I will not comment on whether you have been "scammed" or not. However I guess I would anticipate that the temperatures of the GPU exceeding the CPU in a standard waterloop with a 2070+8700K to be the otherway around.

 

EKWB are usually very liberal when a customer thinks that a product does not meet expected performance and typically will accept RMA claims of this nature (usually replacing the block in question to rule out a faulty individual product) and so this is what I would suggest as a next course of action.

[Flowroro]

1 hour ago, For Science! said:

Since I cannot rule out that the block installation was somehow compromised (i.e. pic of PCB for block installation) I will not comment on whether you have been "scammed" or not. However I guess I would anticipate that the temperatures of the GPU exceeding the CPU in a standard waterloop with a 2070+8700K to be the otherway around.

 

EKWB are usually very liberal when a customer thinks that a product does not meet expected performance and typically will accept RMA claims of this nature (usually replacing the block in question to rule out a faulty individual product) and so this is what I would suggest as a next course of action.

Yea I might try that, temps on a 7700k at 5ghz plus a 1080 oced with a singular 240 rad same equipment gpu managed to only be 52c. True my ambient is a few c higher but still with an extra 360 rad it shouldn’t be higher anyway. I’ll try contact ek about a replacement block thanks for the tip

[Flowroro]

I’ll take a pic of the block after I remove it from the gpu 

[Flowroro]

6 hours ago, fabafaba said:

I think some 1080TIs just have some problem with their temp measurement. My 1080TI is at 26°C right now, but the water going into it is at 29°C, set to a 3°C delta over ambient. I checked all of my temp sensors against a thermocouple for a multimeter and a BBQ thermometer when I upgraded my loop two weeks ago, and they were all within 0.5°C. It seems like some chips/cards just aren't properly calibrated.

 

Are they using the fluid gaming stuff too? I wasn't very happy with my A240G kit and additional 360mm radiator too. I think it just comes down to the weak pump, flimsy little radiators (the "core" is just 15-16mm thick) and the aluminum blocks with a lower heat conductivity compared to copper, as well as fewer and thicker fins in the GPU block.

 

2 hours ago, For Science! said:

Since I cannot rule out that the block installation was somehow compromised (i.e. pic of PCB for block installation) I will not comment on whether you have been "scammed" or not. However I guess I would anticipate that the temperatures of the GPU exceeding the CPU in a standard waterloop with a 2070+8700K to be the otherway around.

 

EKWB are usually very liberal when a customer thinks that a product does not meet expected performance and typically will accept RMA claims of this nature (usually replacing the block in question to rule out a faulty individual product) and so this is what I would suggest as a next course of action.

So I noticed that if I tilt my pc on its back and then back to upright and turn on my pc, the tube that goes from gpu to pump looks like the liquid only flows on the side on the side of the tube and there’s a huge hollow air hole until I shake my pc and then it kinda gets smaller as I shake it. If you zoom into the photo you can kind of see it but it looks as if liquid is struggling to flow out of the gpu block. Could this be the issue? But if this was the case the whole loop would be obstructed resulting in higher cpu temps too. But then 50-55c while just gaming with a stock delidded 8700k seems a tad high too. This is 30-35c over ambient. You can see near the fitting how there’s an inconsistency in fluid it almost looks as if something is obstructing half the flow, but when I tilt it side to side almost flat down each side it will disappear kind of but not sure fully cuz I can’t see the tubing inside the fitting or the flow because the big air bubbles in the block aren’t moving at all. The pumps been running at full speed for quite some time now.

[fabafaba]

14 minutes ago, Flowroro said:

<snip>

That's what I meant when I said the pump is weak. This does however only affect your GPU temps. While the water gets hotter as it flows through the blocks because it spends more time in there, it also spends more time in the radiators to get cooler, so the water temperature behind the rads stays pretty much the same, regardless of the flowrate. I usually have a temperature difference before and after the blocks of about 2-3°C and actually use this difference to regulate the pump speed. But by putting a lot of heat into the system and turning the pump speed manually down, I can force a difference of 7-8°C with the same water temperature behind the radiators at the same fanspeeds.

But whether or not the flow rate is actually the issue in your loop, I'd recommend you get a D5 or DDC pump to get a decent flow rate. If you end up going for a copper loop, you'll already have a good (and quieter) pump.

[Flowroro]

22 minutes ago, fabafaba said:

That's what I meant when I said the pump is weak. This does however only affect your GPU temps. While the water gets hotter as it flows through the blocks because it spends more time in there, it also spends more time in the radiators to get cooler, so the water temperature behind the rads stays pretty much the same, regardless of the flowrate. I usually have a temperature difference before and after the blocks of about 2-3°C and actually use this difference to regulate the pump speed. But by putting a lot of heat into the system and turning the pump speed manually down, I can force a difference of 7-8°C with the same water temperature behind the radiators at the same fanspeeds.

But whether or not the flow rate is actually the issue in your loop, I'd recommend you get a D5 or DDC pump to get a decent flow rate. If you end up going for a copper loop, you'll already have a good (and quieter) pump.

So if I reversed the flow order, going from pump into gpu directly, giving it less flow loss would that mean I’d get higher cpu temps and lower gpu? Theoretically speaking. I know generally order doesn’t affect temps but since there looks like a flow rate drop as it goes through the loop...

[Flowroro]

23 hours ago, For Science! said:

This is fine, I am therefore going to assume your idle water temperature is 28 degrees. Let it climb until the CPU/GPU temp is stablized. Given you have probably less than 1 liter of fluid, it should take between 30-60 minutes. 

 

Also, what was your GPU cooler model (i.e. what is it a founders edition? a Strix? a FTW3).

so leaving my game running for 10 hours. gpu is 30-40c over ambient, and cpu is 20c over ambient. this still seems too high

[fabafaba]

9 hours ago, Flowroro said:

So if I reversed the flow order, going from pump into gpu directly, giving it less flow loss would that mean I’d get higher cpu temps and lower gpu? Theoretically speaking. I know generally order doesn’t affect temps but since there looks like a flow rate drop as it goes through the loop...

The flow rate is the same at any point in the loop, and you won't change the flow by changing the loop order. It will however affect the temperature of the components, and the second block in the loop always gets warmer water. The difference in temperature is usually 1-2°C with a sufficient flow rate and doesn't change the cooling performance. But if you get barely any fluid moving, you'll see significantly higher temps on the second component in the loop.

3 hours ago, Flowroro said:

so leaving my game running for 10 hours. gpu is 25-26c over ambient, and cpu is 20c over ambient. this still seems too high

I don't think there's anything you can do besides getting a stronger pump. The problem with slim fans like the aluminum ones is that they barely move any heat when the water is just slightly above ambient. Radiators become more efficient as the difference gets bigger. I only learned about this a couple of days ago and I'm currently playing around a little with my pc. When in idle, I used to need to run my fans at 900-1000rpm to hold a 3°C delta over ambient. But when I allow a higher difference, I just set it to 10°C for now, my fans run at their minimum of 450rmp and actually hold 9°C. My best gueass by now is that the fluid gaming kits just run a little hotter, both because the aluminum blocks just not as good as copper blocks and because the radiators need to run hot to be more efficient. EKWB explains it in this blogpost here:

https://www.ekwb.com/blog/radiators-part-2-performance/

And if the aluminum rads need a 10-15°C difference between the water and ambient, that's how you get your temps.

[Flowroro]

19 minutes ago, fabafaba said:

The flow rate is the same at any point in the loop, and you won't change the flow by changing the loop order. It will however affect the temperature of the components, and the second block in the loop always gets warmer water. The difference in temperature is usually 1-2°C with a sufficient flow rate and doesn't change the cooling performance. But if you get barely any fluid moving, you'll see significantly higher temps on the second component in the loop.

I don't think there's anything you can do besides getting a stronger pump. The problem with slim fans like the aluminum ones is that they barely move any heat when the water is just slightly above ambient. Radiators become more efficient as the difference gets bigger. I only learned about this a couple of days ago and I'm currently playing around a little with my pc. When in idle, I used to need to run my fans at 900-1000rpm to hold a 3°C delta over ambient. But when I allow a higher difference, I just set it to 10°C for now, my fans run at their minimum of 450rmp and actually hold 9°C. My best gueass by now is that the fluid gaming kits just run a little hotter, both because the aluminum blocks just not as good as copper blocks and because the radiators need to run hot to be more efficient. EKWB explains it in this blogpost here:

https://www.ekwb.com/blog/radiators-part-2-performance/

And if the aluminum rads need a 10-15°C difference between the water and ambient, that's how you get your temps.

well what a waste of money. all the reviews have way better temps. i still cant believe the temps are so atrocious. these temps are actually disgusting and if its "normal" ek is basically scamming people... can spend 200 more and get a real kit which apparently will yield 20c difference under load which also doesnt sound right, should be a few c max... im staring at my gpu pcb right now and the section where the block flow ports are the pcb is slightly bent upwards. is this normal? perhaps ive put thermal pads where i shouldnt have and causing this issue? i still have hope

[Flowroro]

14 hours ago, For Science! said:

Since I cannot rule out that the block installation was somehow compromised (i.e. pic of PCB for block installation) I will not comment on whether you have been "scammed" or not. However I guess I would anticipate that the temperatures of the GPU exceeding the CPU in a standard waterloop with a 2070+8700K to be the otherway around.

 

EKWB are usually very liberal when a customer thinks that a product does not meet expected performance and typically will accept RMA claims of this nature (usually replacing the block in question to rule out a faulty individual product) and so this is what I would suggest as a next course of action.

taking another closer look at your gpu block, the contact doesnt seem to be as good as it looks where it presses so hard against the gpu that it squishes out all the paste. will reseat it again, take off some thermal pads that according to the same pcb diagram of a different gpu that uses the same block that shouldnt be there, and check thermal paste coverage again.

[fabafaba]

2 minutes ago, Flowroro said:

im staring at my gpu pcb right now and the section where the block flow ports are the pcb is slightly bent upwards. is this normal? perhaps ive put thermal pads where i shouldnt have and causing this issue? i still have hope

That sounds like too much mounting pressure on the GPU die, not incorrect thermal pad application.

You can see on this block, how the portion that touches the die is slightly proud of the "base level" of the block. If you tighten the screws around there too much, you pull the PCB down around the die, which in turn gets pushed up, actually decreassing the mounting pressure in the middle of the die and possibly seperating the TIM. You also run the risk of seperating the layers of the PCB when it bends too much, cracking the solderjoints under the GPU die or even cracking the die when the pressure on the edges of the die is too high and the pressure in the middle is too low. When installing a GPU block you'll have to pay close attention to make sure the PCB doesn't bend and is parallel to the edge of the block. Just tighten the screws until you feel a resistance, then check for bends and if there aren't any, go a little tighter and check again. Once you see a slight bend, you'll want to back the screws off by a quarter turn or so.

[For Science!]

4 minutes ago, Flowroro said:

taking another closer look at your gpu block, the contact doesnt seem to be as good as it looks where it presses so hard against the gpu that it squishes out all the paste. will reseat it again, take off some thermal pads that according to the same pcb diagram of a different gpu that uses the same block that shouldnt be there, and check thermal paste coverage again.

Please please snap a pic and post it. Makes it much easier for me to tell if anything obvious is wrong.

[Flowroro]

22 minutes ago, fabafaba said:

That sounds like too much mounting pressure on the GPU die, not incorrect thermal pad application.

You can see on this block, how the portion that touches the die is slightly proud of the "base level" of the block. If you tighten the screws around there too much, you pull the PCB down around the die, which in turn gets pushed up, actually decreassing the mounting pressure in the middle of the die and possibly seperating the TIM. You also run the risk of seperating the layers of the PCB when it bends too much, cracking the solderjoints under the GPU die or even cracking the die when the pressure on the edges of the die is too high and the pressure in the middle is too low. When installing a GPU block you'll have to pay close attention to make sure the PCB doesn't bend and is parallel to the edge of the block. Just tighten the screws until you feel a resistance, then check for bends and if there aren't any, go a little tighter and check again. Once you see a slight bend, you'll want to back the screws off by a quarter turn or so.

so i just untightened the screws and screwed them back in until i feel a little bit of resistance. the pcb is now slightly less bent, minimal bend left. temperatures are worse already. its not like i overtighten them, i initially screwed them in until light turning of the screw no longer will move. is this too much?

[fabafaba]

1 minute ago, Flowroro said:

so i just untightened the screws and screwed them back in until i feel a little bit of resistance. the pcb is now slightly less bent, minimal bend left. temperatures are worse already. its not like i overtighten them, i initially screwed them in until light turning of the screw no longer will move. is this too much?

I think at this point we really need a pic

[Flowroro]

You’ll have to zoom in very close to see the bend. It’s very subtle but it’s there. 

2 minutes ago, fabafaba said:

I think at this point we really need a pic

 

[Flowroro]

This is with hand tight screwed in. 

[For Science!]

Since I found it while browsing the other sub forums. OP has put at least 2 thermal pads where there shouldn’t be. This is probably the reason, highlights the importance of posting photos.

 

https://www.ekfluidgaming.com/EK-IM/EK-IM-3831109817049.pdf

 

3 hours ago, Flowroro said:

The block didnt come with an instructions manual so based on my markings on the image below is there a place where I shouldnt/dont need to put thermal pads? having some issues with gpu temps wondering if gpu is not making good contact because of this.

 

[Flowroro]

1 hour ago, For Science! said:

Since I found it while browsing the other sub forums. OP has put at least 2 thermal pads where there shouldn’t be. This is probably the reason, highlights the importance of posting photos.

 

https://www.ekfluidgaming.com/EK-IM/EK-IM-3831109817049.pdf

 

 

Ok so update I removed all thermal pads that weren’t on that document. Gpu is still bent why is this???

[For Science!]

14 minutes ago, Flowroro said:

Ok so update I removed all thermal pads that weren’t on that document. Gpu is still bent why is this???

I personally dont worrk about board flex so much. A picture of your now updated pad placement would be appreciated.

[Flowroro]

25 minutes ago, For Science! said:

I personally dont worrk about board flex so much. A picture of your now updated pad placement would be appreciated.

Heres the updated pad placement. removed quite a few

[For Science!]

11 minutes ago, Flowroro said:

Heres the updated pad placement. removed quite a few

Make sure you remove the pad on the right side as well close to the power plugs. A photo of the actual card with the actual thermal pads right before you put the block on is desirable instead of a mock up.