Gtx 1080 to strix getting too hot for water cooling?

[Dannylti95]

15 minutes ago, For Science! said:

I used Arctic MX-4, some people had an opinion of Arctic Silver 5 being capacitative, But any good normal paste is good, So I in fact have a mish-mash of pastes:

 

CPU--> IHS = Conductonaut

IHS --> Cooler = Arctic Silver 5

GPU --> Cooler = Arctic MX4

Lol I'm using the same m 4 for my blocks so I'm good there I'll look into condutctonaut for the ihs.

[Dannylti95]

9 minutes ago, For Science! said:

Seconded, I have 2x 1080 using bitspower blocks and my temps are fine,

Nice! do you think the thermal pads that go on the backplate matter much? I installed them and the backplate gets quite warm. At this point Im really puzzled. I spent nearly 2 months doing research, watching jayz2cents on YouTube (and well other YouTubers too but mostly jay) and nonstop planning. All to get this problem:/

[For Science!]

Just now, Dannylti95 said:

-snip-

Definitely leave them on, one way to think about it is that if the backplate is hot, that is temperature from the chips being dissipated using the increased surface area. My backplate is hot too, but I also know that the chips they sit under used to be even hotter,

[Dannylti95]

29 minutes ago, For Science! said:

Definitely leave them on, one way to think about it is that if the backplate is hot, that is temperature from the chips being dissipated using the increased surface area. My backplate is hot too, but I also know that the chips they sit under used to be even hotter,

What are your temps on the 1080s?

[For Science!]

3 minutes ago, Dannylti95 said:

What are your temps on the 1080s?

What test would you like me to run?

[Dannylti95]

10 minutes ago, For Science! said:

What test would you like me to run?

Temps under load and idle if u don't mind, are u on a single or dual loop?

[For Science!]

20 minutes ago, Dannylti95 said:

Temps under load and idle if u don't mind, are u on a single or dual loop?

I used Realbench Stress Test (version 2.54) if you want to compare. Unfortuantely I don't have time to run it for a long periods of time at the moment, but you can add maybe 5~10 degrees as the water warms up (I am more silence orientated so my fans are very low rpm ~800 rpm). I have a single loop in a parallel configuration for the GPU.

 

On Idle I am currently low 40s for the CPU and high 30s for the GPU, I am literally in the process in cooking a gigantic curry in the same room and my ambient temperature is pretty hot.

Spoiler

After firing up realbench, my temps are close to 80 for the CPU (overclocked to 4.8 GHz) and close to 50 for GPU (also overclocked). My system has been on for a while so the water had equilibriated with idle temps. If I were to let this continue to run for a few hours, I would expect my CPU to be in the mid-high 80's and the GPU to be mid-high 50's. Obviously fan speed depending (as well as ambient)

 

Spoiler

For your reference, this is my setup. Note that I have VRMs in my loop too, so even more heat being dumped into the system

Spoiler

 

[rasmuskrj]

52 minutes ago, Dannylti95 said:

Nice! do you think the thermal pads that go on the backplate matter much? I installed them and the backplate gets quite warm. At this point Im really puzzled. I spent nearly 2 months doing research, watching jayz2cents on YouTube (and well other YouTubers too but mostly jay) and nonstop planning. All to get this problem:/

I get it man. I also spent months researching before building my loop.

 

As I see it there are two general things that could explain this:

 

1. The heat is not transfered effeciently from the GPU to the waterblock. If this is that case, then your water temps should be fairly low. You can buy a temperature probe or simply put a hand on the tubes or rads to get an idea. It could also be a massive air bubble or something inside the GPU block, preventing transfer of heat from the block to the rest of the system. Another possible explaination could be that the block is is mounted incorrectly or that it is indeed faulty somehow(although I find this unlikely).
2. The radiators are not able to dissipate the heat in the water. In this case the water would be very hot, again something you should get an idea of by touching tubes and radiators. This could again be caused by air bubbles in the radiators. Other explainations could be that water is not flowing correctly because of a kink or maybe a defective pump. How loud is the pump when you put it at max?

Of course there is the off chance that the temperature sensor on the GPU is giving you wrong readings, but I don't think I've heard of anyone experiencing this.

[For Science!]

2 hours ago, Dannylti95 said:

-snip-

On another unrelated note, did you forget to peel off the protective sticker off the anticyclone EK logo thing (that you have as decoration I assume) since there will not be any cycloning up there =P

[Dannylti95]

7 hours ago, For Science! said:

I used Realbench Stress Test (version 2.54) if you want to compare. Unfortuantely I don't have time to run it for a long periods of time at the moment, but you can add maybe 5~10 degrees as the water warms up (I am more silence orientated so my fans are very low rpm ~800 rpm). I have a single loop in a parallel configuration for the GPU.

 

On Idle I am currently low 40s for the CPU and high 30s for the GPU, I am literally in the process in cooking a gigantic curry in the same room and my ambient temperature is pretty hot.

  Hide contents

After firing up realbench, my temps are close to 80 for the CPU (overclocked to 4.8 GHz) and close to 50 for GPU (also overclocked). My system has been on for a while so the water had equilibriated with idle temps. If I were to let this continue to run for a few hours, I would expect my CPU to be in the mid-high 80's and the GPU to be mid-high 50's. Obviously fan speed depending (as well as ambient)

 

  Hide contents

For your reference, this is my setup. Note that I have VRMs in my loop too, so even more heat being dumped into the system

  Hide contents

 

Gorgeous rig man. And wow those temps look really in check. I'm doing some shopping right now, what do you think I should do? Either but a crapton of 90° bends so I don't kink lines and do a single loop again( I gotta re do it either way didn't add a drain valve and it's a major pain in the ass to drain) or do dual loop?

 

5 hours ago, For Science! said:

On another unrelated note, did you forget to peel off the protective sticker off the anticyclone EK logo thing (that you have as decoration I assume) since there will not be any cycloning up there =P

I did forget to peel it off lol but I think it looks cool with the white sticker I might leave it

 

7 hours ago, rasmuskrj said:

I get it man. I also spent months researching before building my loop.

 

As I see it there are two general things that could explain this:

 

1. The heat is not transfered effeciently from the GPU to the waterblock. If this is that case, then your water temps should be fairly low. You can buy a temperature probe or simply put a hand on the tubes or rads to get an idea. It could also be a massive air bubble or something inside the GPU block, preventing transfer of heat from the block to the rest of the system. Another possible explaination could be that the block is is mounted incorrectly or that it is indeed faulty somehow(although I find this unlikely).
2. The radiators are not able to dissipate the heat in the water. In this case the water would be very hot, again something you should get an idea of by touching tubes and radiators. This could again be caused by air bubbles in the radiators. Other explainations could be that water is not flowing correctly because of a kink or maybe a defective pump. How loud is the pump when you put it at max?

Of course there is the off chance that the temperature sensor on the GPU is giving you wrong readings, but I don't think I've heard of anyone experiencing this.

Yeah I've been thinking about all that trying to be real sciency plus doing thermodynamics research on computers lol. STILL CAN FIGURE IT OUT :((((((((

[For Science!]

24 minutes ago, Dannylti95 said:

-snip-

If I were you I wouldn't do a dual loop unless you are after specific looks, or if you are likely to swap out one component more regularly than another, or if you have a crazy restrictive system (like quad-SLI). Drain valve is a big must in my opinion too.

 

I would just try to re-do the bends with a proper tight fitting silicon insert and just practise till perfection. My bends are hand done (no mandrels) and I also over bent some tubes in the beginning. Fittings are a good way to eliminate some super tight bends around the rads for example, but I wouldnt do an all fitting bend system (I don't like the looks and its more restrictive as well).

 

I would also peel off that sticker, who knows what residues it may have and affect the lifetime of the coolant/

[Dannylti95]

2 hours ago, For Science! said:

If I were you I wouldn't do a dual loop unless you are after specific looks, or if you are likely to swap out one component more regularly than another, or if you have a crazy restrictive system (like quad-SLI). Drain valve is a big must in my opinion too.

 

I would just try to re-do the bends with a proper tight fitting silicon insert and just practise till perfection. My bends are hand done (no mandrels) and I also over bent some tubes in the beginning. Fittings are a good way to eliminate some super tight bends around the rads for example, but I wouldnt do an all fitting bend system (I don't like the looks and its more restrictive as well).

 

I would also peel off that sticker, who knows what residues it may have and affect the lifetime of the coolant/

Ok, I mean I do wanna go for the dual loop just cause it'll fill up the case a bit more, I might add another 1080 ti in there as well so I wanna have a dedicated loop. Yes, I completely agree on the drain valve, and these bends were all made by hand, took me about 2 days and 2 sets of uv tubing lol. I'll add some 90° fittings then but not too many, and I'll make sure to peel that sticker off as well you're right about the residue, might become a problem later on. I also should mention I was using a 650 watt psu. Now I just swapped it out with a 750 watt and GPU temps dropped down to 68° C max. Idk if that really makes a big difference but I did get a 4 degree drop, ambient temp by me is about 80° f but I have the ac on almost all times so that always plays a factor. 

[Vigilante505]

You probably thought of this already but I was having a similar issue. On my loop I was getting 30C idle and 75C full load and thought it was a bit high. I removed the rad and noticed a giant air bubble pop out of it. Shaking the case didnt get it out when I installed it.

 

I dropped my temps a bit further afterwards by changing the fans around.

[Dannylti95]

42 minutes ago, Vigilante505 said:

You probably thought of this already but I was having a similar issue. On my loop I was getting 30C idle and 75C full load and thought it was a bit high. I removed the rad and noticed a giant air bubble pop out of it. Shaking the case didnt get it out when I installed it.

 

I dropped my temps a bit further afterwards by changing the fans around.

I shook the case for quite a bit already but I'll still give it a try just for extra measure, I really hope that's all it is lol

[For Science!]

3 hours ago, Dannylti95 said:

I shook the case for quite a bit already but I'll still give it a try just for extra measure, I really hope that's all it is lol

You really need to rock it, lay it flat down on its back etc (move it on the four courners individually), if you can lift it, rotate it in the air as well - while the pump is running. This is really part of the bleeding process.

[Dannylti95]

3 hours ago, For Science! said:

You really need to rock it, lay it flat down on its back etc (move it on the four courners individually), if you can lift it, rotate it in the air as well - while the pump is running. This is really part of the bleeding process.

Did all that:P

[Dannylti95]

What 90° double compression fittings would you guys recommend?

[For Science!]

Bitspower maybe? If you have metric exact tubing, that is.

https://www.bitspower.com.tw/index.php?main_page=product_info&cPath=155_164&products_id=3417&zenid=os4ifom3s37e9gi876h60vk830

 

[Dannylti95]

4 hours ago, For Science! said:

Bitspower maybe? If you have metric exact tubing, that is.

https://www.bitspower.com.tw/index.php?main_page=product_info&cPath=155_164&products_id=3417&zenid=os4ifom3s37e9gi876h60vk830

 

Not really sure, I have primochill rigid 1/2 od 3/8 I'd tubing

[For Science!]

The only other thing I just thought is whether you put enough paste on your GPU die, Bitspower does not have any recommendation on their instructions so I suppose there can be quite a lot of variation between people. I went with EKWB style, see photo below; as we know from LTT Workshops, application method does not matter except for when you apply too little.

 

 

Other pics

Spoiler

31 minutes ago, Dannylti95 said:

-snip-

Sounds like you might be out of luck, I only know bitspower doing angled fittings with compression fittings pre-attaached.

[For Science!]

4 minutes ago, For Science! said:

-snip-

P.S. The heavy duty wrench played no role in the assembly of my watercooled PC

[Dannylti95]

5 hours ago, For Science! said:

The only other thing I just thought is whether you put enough paste on your GPU die, Bitspower does not have any recommendation on their instructions so I suppose there can be quite a lot of variation between people. I went with EKWB style, see photo below; as we know from LTT Workshops, application method does not matter except for when you apply too little.

 

 

Other pics

  Hide contents

Sounds like you might be out of luck, I only know bitspower doing angled fittings with compression fittings pre-attaached.

I did put quite a bit of paste on it the ek method as well. Now I'm checking out the pump in my rig, fans and everything else is off except the pump. I feel vibration on the tube coming out of it but not so much on the pump. I do hear a humming sound but I think that's just the pump running? Idk it's not loud at all

[For Science!]

1 minute ago, Dannylti95 said:

-snip-

sounds normal to me, the D5 is quite a quiet pump; just like you I feel the most vibration in the tube coming out of the pump rather than the pump itself. As others have eluded to, you can quantitatively put your flow rate questions if you invest in a flowmeter + multiplier, but I think it looks hideous and is not so useful in the long run (I used to have one in my system, but removed it after establishing I had enough flow).

[Dannylti95]

30 minutes ago, For Science! said:

sounds normal to me, the D5 is quite a quiet pump; just like you I feel the most vibration in the tube coming out of the pump rather than the pump itself. As others have eluded to, you can quantitatively put your flow rate questions if you invest in a flowmeter + multiplier, but I think it looks hideous and is not so useful in the long run (I used to have one in my system, but removed it after establishing I had enough flow).

Ok sounds good to me then, I don't think I need a flow meter, in fact I shook the living crap out of my case in all directions (making sure reservoir air doesn't get into the system) and a crap ton of bubbles came out of the 480 so I'm gonna shake it around for a couple of hours the fire it up, fingers crossed this is my problem

[Dannylti95]

Well bad news, even after more air bleeding, I m still getting really bad temps. I'm gonna take the block apart, this is my 3rd time-_-