Should I change order on Water loop?

[Livinloud]

Currently I have res>Gpu>Gpu>140mm rad>cpu>240rad>res again.

 

I was wondering if I should reverse the order? my Gpu's top at 50c where as my i9-9900 tops at 93c when I use folding at home. Normally when I don't use gpu's at max cpu reaches 85c tops.

[W-L]

3 minutes ago, Livinloud said:

-SNIP-

The order of a loop does make a difference in terms of performance due to how fast the fluid moves around the loop. As for achieving better temps especially if you are maxing out your GPU's you need more rad space, you are very much at the absolute bare minimum of 120mm of space per component. A good measure is usually 120mm per component plus 120-240mm for extra, more if you want dead silence. I would recommend something like 4-5 x120mm of total rad space. 

[Livinloud]

1 minute ago, W-L said:

The order of a loop does make a difference in terms of performance due to how fast the fluid moves around the loop. As for achieving better temps especially if you are maxing out your GPU's you need more rad space, you are very much at the absolute bare minimum of 120mm of space per component. A good measure is usually 120mm per component plus 120-240mm for extra, more if you want dead silence. I would recommend something like 4-5 x120mm of total rad space.

well issue I have is I could possibly put another rad in on top but then I am messing with clearance from mobo. Right now mow 140mm fans are pushed up by mobo. Plus my system is all aluminum so finding aluminum rads isn't easy.

[Spotty]

I agree with @W-L. 140mm rad + 240mm rad for an overclocked 9900k and two 2080Tis is going to be pushing it. I don't think loop order is going to matter that much. What case do you have? Any option to add another 240/280mm rad?

[W-L]

2 minutes ago, Livinloud said:

well issue I have is I could possibly put another rad in on top but then I am messing with clearance from mobo. Right now mow 140mm fans are pushed up by mobo. Plus my system is all aluminum so finding aluminum rads isn't easy.

If you have the fluid gaming series from EKWB they have aluminum components for expanding the loop. If you can put in another dual rad a 240/280mm it would help greatly.

https://www.ekfluidgaming.com/parts/

[Livinloud]

1 minute ago, Spotty said:

I agree with @W-L. 140mm rad + 240mm rad for an overclocked 9900k and two 2080Tis is going to be pushing it. I don't think loop order is going to matter that much. What case do you have? Any option to add another 240/280mm rad?

Phanteks P400s.

[Livinloud]

After some youtube videos of overclocking I adjusted some voltages and messed around with some of the core speeds and running 80c max on prime95 high temp test.

 

Running all cores at 5.0Ghz cinebench score 2107 just like Deabaur (spelling?? LOL) temps reached 90C.

 

Running 2 cores at 5.1Ghz and 2 at 5.0Ghz and rest at 4.9Ghz getting 80C and cinebench score of 2100. I think this is a big improvement. I am running Vcore at 1.36 with a little give so max it does reach 1.38V.

 

If you think it is high let me know. I am still a novice Overclocker. But just running Cinebench it reaches 70C. All other high temps were with prime95. If I run blend test on prime I get high of 63C.

 

Motherboard is a z370-H Asus. I think if I had a newer board I could run at a lower vcore but I could be wrong.

 

[Benjeh]

On 2/17/2019 at 3:45 AM, Livinloud said:

After some youtube videos of overclocking I adjusted some voltages and messed around with some of the core speeds and running 80c max on prime95 high temp test.

 

Running all cores at 5.0Ghz cinebench score 2107 just like Deabaur (spelling?? LOL) temps reached 90C.

 

Running 2 cores at 5.1Ghz and 2 at 5.0Ghz and rest at 4.9Ghz getting 80C and cinebench score of 2100. I think this is a big improvement. I am running Vcore at 1.36 with a little give so max it does reach 1.38V.

 

If you think it is high let me know. I am still a novice Overclocker. But just running Cinebench it reaches 70C. All other high temps were with prime95. If I run blend test on prime I get high of 63C.

 

Motherboard is a z370-H Asus. I think if I had a newer board I could run at a lower vcore but I could be wrong.

 

80c+ on water is not good. I'm running a 6700k I know its a cooler chip but at full tilt with 1.45v pumped into it, max temp i see is 70c and that's a spike, constant load will see me around 50c. You need more rad space like suggested, you're loop is restrictive with the 3 waterblocks so maybe a better pump, would be a worthy investment I'm still using a D5 I've had for 4+ years now.

[Livinloud]

On 2/18/2019 at 1:02 PM, Benjeh said:

80c+ on water is not good. I'm running a 6700k I know its a cooler chip but at full tilt with 1.45v pumped into it, max temp i see is 70c and that's a spike, constant load will see me around 50c. You need more rad space like suggested, you're loop is restrictive with the 3 waterblocks so maybe a better pump, would be a worthy investment I'm still using a D5 I've had for 4+ years now.

Well after some long thought I'm gonna sell one of my gpu's. Also probably upgrading to a new mobo..looking at gigabyte due to I am getting vrm throttling as well. Never noticed till I ran prime 95 for awhile. Most benchmarks are done at min or two.

[Benjeh]

36 minutes ago, Livinloud said:

Well after some long thought I'm gonna sell one of my gpu's. Also probably upgrading to a new mobo..looking at gigabyte due to I am getting vrm throttling as well. Never noticed till I ran prime 95 for awhile. Most benchmarks are done at min or two.

Why would changing motherboard stop your throttling? Is it well documented that certain boards have poor VRM? Also you do know that a waterblock on your VRM will solve the thermal throttling as we used to do that with the 990FX chip set for 8350's overclocked to 5ghz. You can also place a fan directly over the VRM heatsink and put a fan behind the CPU socket, this will also help reduce temps.

[Benjeh]

Found a pic of my old setup, I cooled the VRM and the NB
 

Spoiler

 

[Livinloud]

26 minutes ago, Benjeh said:

Why would changing motherboard stop your throttling? Is it well documented that certain boards have poor VRM? Also you do know that a waterblock on your VRM will solve the thermal throttling as we used to do that with the 990FX chip set for 8350's overclocked to 5ghz. You can also place a fan directly over the VRM heatsink and put a fan behind the CPU socket, this will also help reduce temps.

Having a higher phase count, they can distribute the power so creating less heat. Waterblock on the VRM's? I have seen it before but since my system is aluminum I don't believe they make that in that alloy. I do have plenty of air flow in the case, but I suppose placing a fan over them would help with the thermals. Now for looks not sure how to pull that off.

[Mick Naughty]

Need to be watching the fluid temps as well. If you aren’t already. Imagine they are dangerously high given the temps stated. 

[Livinloud]

On Wednesday, February 20, 2019 at 8:13 AM, Mick Naughty said:

Need to be watching the fluid temps as well. If you aren’t already. Imagine they are dangerously high given the temps stated.

Idk if water temp is bad because once load is done it drops down to 25c but I have since upgraded my motherboard to a Gigabyte z390 ultra. My only issue right now is my vcore has a max of 1.78c (holy cow that's high to me). But I have only changed the XMP profile and a few cores to 5.0 on first boot (not wise I know but it worked). When I get home I need to mess around in the bios..but I am not familiar with gigabyte at all..Asus is by far the best bios but switching boards so far as been night and day.

 

[Mick Naughty]

5 minutes ago, Livinloud said:

Idk if water temp is bad because once load is done it drops down to 25c but I have since upgraded my motherboard to a Gigabyte z390 ultra. My only issue right now is my vcore has a max of 1.78c (holy cow that's high to me). But I have only changed the XMP profile and a few cores to 5.0 on first boot (not wise I know but it worked). When I get home I need to mess around in the bios..but I am not familiar with gigabyte at all..Asus is by far the best bios but switching boards so far as been night and day.

 

Give them be temps of the hardware and the cooling they have, I’d assume the water is high. The board doesn’t support temp sensors?

[Livinloud]

17 minutes ago, Mick Naughty said:

Give them be temps of the hardware and the cooling they have, I’d assume the water is high. The board doesn’t support temp sensors?

It came with 2 thermo's but not sure where to put them at right now. I have only been able to update all the software for the new motherboard. Last step is flashing the bios to the most updated but I may hold off as it is work well for me. On the vcore the program that showed it max at 1.78v was OCCT. I have since remotely accessed my computer and used another monitoring program and ran the same programs and it says make vcore was 1.26v. But only thing that looks remotely close was VIN7 at 1.7v so what ever VIN7 is.

[xreaperx22]

On 2/8/2019 at 9:43 PM, W-L said:

The order of a loop does make a difference in terms of performance due to how fast the fluid moves around the loop. As for achieving better temps especially if you are maxing out your GPU's you need more rad space, you are very much at the absolute bare minimum of 120mm of space per component. A good measure is usually 120mm per component plus 120-240mm for extra, more if you want dead silence. I would recommend something like 4-5 x120mm of total rad space. 

doesnt make much difference actually prolly a .05c difference here is a great read on water cooling.

 

http://forums.procooling.com/vbb/showthread.php?t=10301

[W-L]

2 hours ago, xreaperx22 said:

doesnt make much difference actually prolly a .05c difference here is a great read on water cooling.

 

http://forums.procooling.com/vbb/showthread.php?t=10301

In real work performance it's with margin of error so it's a measurable and calculable amount with the right equipment but at the end of the day there's no real difference. 

[Blai5e]

If you're heating up to 90°C and then quickly cooling down to 25°C, I'd lay odds your cold plate contact and/or the thermal paste or isn't right.

[Benjeh]

1 hour ago, Blai5e said:

If you're heating up to 90°C and then quickly cooling down to 25°C, I'd lay odds your cold plate contact and/or the thermal paste or isn't right.

or it needs delidding.

[Livinloud]

Lapped cpu and switched out some tubes so make sure they weren't bending to far so it doesn't kink. Also used different thermal paste and now my thermals are a lot better now. I unfortunately have a chip that can't be stable at 1.28v at 4.9ghz with offset 2. I need to change my LLC back to turbo or feed it some more volts. 1.29v I think it may be stable. But if I wanna reach 5.0 I am going to have to set the Avx to 3 or maybe even 4 to keep thermals down. cinebench score at 4.9 and 1.28v is 2040 (not sure if that is inline or not). But 5.0ghz when I tested at 1.31v got 2177. But using Realbench it crashes on both. So I am booting up and somewhat stable so somewhere in between those settings and messing with LLC I am just gonna go for 4.9ghz.