Should I cancel my liquid metal thermal solution order?

[JonS77]

I've just order a  thermal grizzly conductonaut liquid metal but my cooler is a Syntch Mugen 5. I'm not sure if the conductive material works with the plate of my cooler. Is it safe to use? what do you think? should I cancel the order?

 

CPU: 9900K

 

[JonS77]

7 minutes ago, Genwyn said:

Scratch what I said, old models are aluminum base. The new models are nickel plated copper which are fine with liquid metal.

 

If you have an old model, pre AM4, then don’t. If you ordered yours within the last 2 years then you’re alright.

You can tell the difference by if the contact plate is shiny or matte. Shiny is nickel plated, matte is aluminum.

Thanks, that's very helpful. I've bought it a week ago and according to the product page it should be Nickel-plated copper. At this point I'm also not sure if it is worth the trouble overall and how temp difference it will give me. Considering that there seems to be some risk involved with the application process I've canceled the order to research it further.

 

Do you thin it is worth it? My temps go as high as 85 while on 100% load while 3d rendering.

[Sorenson]

I would say that 85c is where I would start to be concerned. You can try replacing your thermal paste with better quality non-conductive thermal paste, increasing your fan curve, or try lowering your voltage a little bit.

[JonS77]

10 minutes ago, Sorenson said:

I would say that 85c is where I would start to be concerned. You can try replacing your thermal paste with better quality non-conductive thermal paste, increasing your fan curve, or try lowering your voltage a little bit.

I've bought the 9900k thinking that I could use it at 5GHz. Right now my base is 4.3GHz and my Turbo is at 4.7GHz and my temps still get too high. What are the disadvantages of lowering my voltage? how much performance impact would it have?

 

Right now 5GHz is just impossible with these temps.

[WoodenMarker]

8 minutes ago, Jon-Slow said:

I've bought the 9900k thinking that I could use it at 5GHz. Right now my base is 4.3GHz and my Turbo is at 4.7GHz and my temps still get too high. What are the disadvantages of lowering my voltage? how much performance impact would it have?

What are your temps and voltage at? 

Lower voltage may result in lower stability and lower temps. It shouldn't have any effect on performance as long as it's stable. 

There isn't much benefit to using liquid metal over good thermal paste between the IHS and cooler. I'd recommend sticking with the stock thermal paste that came with the cooler. 

[Sorenson]

Lowering the voltage just reduces the power the chip is receiving. Every CPU has a tolerance to how much voltage you can reduce it by and get the same performance with lower heat produced. It can take a good amount of effort to get the voltage correct depending on your motherboard, and how easy it is to adjust. I would say use this as a last resort if you are unwilling to just buy a better cpu cooler or better thermal paste.

[JonS77]

2 hours ago, WoodenMarker said:

What are your temps and voltage at? 

Lower voltage may result in lower stability and lower temps. It shouldn't have any effect on performance as long as it's stable. 

There isn't much benefit to using liquid metal over good thermal paste between the IHS and cooler. I'd recommend sticking with the stock thermal paste that came with the cooler. 

I had the XMP on and the CPU was set on 1.35v. I turned it off and it is now at 1.243v. But my performance did take a hit. I'm using AC Odyssey benchmark tool. I guess maybe I should have gone with the Ryzen 3900x while making a decision as this processor can't practically hit 5GHz( at least for me)

[JonS77]

1 hour ago, Sorenson said:

Lowering the voltage just reduces the power the chip is receiving. Every CPU has a tolerance to how much voltage you can reduce it by and get the same performance with lower heat produced. It can take a good amount of effort to get the voltage correct depending on your motherboard, and how easy it is to adjust. I would say use this as a last resort if you are unwilling to just buy a better cpu cooler or better thermal paste.

 

I am willing to buy new stuff. For the CPU cooler I did a lot of research and Scythe Mugen 5 rev.b  seemed to on the same level of performance as the Noctua NH-D15, only slightly cheaper. I don't know if there would have been much difference unless NH-D15 is noticeably better. As for the Thermal paste, I am looking into it but I'm not sure how much of an improvement that is. currently, was looking into buying a 'thermal grizzly conductonaut liquid metal' but I'm not sure how much better it would be compared to the 'thermal grizzly kryonaut' or my current Scythe Mugen stock thermal solution or my 2 years old 'Cooler Master E2 IC Essential'.

 

I am also willing to look into liquid or water cooling but I'm not sure if they can give me any advantages since most reviews suggests I should stick with the Air cooling.

[Sorenson]

Thermal paste should ideally be replaced every two years, as it becomes less effective over time. I would recommend either thermal grizzly cryonaut or noctua NT-H1. Noctua coolers are considered the best of the best in terms of silence and cooling but are kind of ugly... unless 1960’s winnebagos are your thing. The bequiet dark rock series are also pretty highly rated. 

 

As far as water cooling, it works really well but costs more. If you move your PC around a lot it is nice because there is not much weight on the CPU socket. If you like aesthetics i personally think water cooling is the way to go, but you might have to cable manage well as all the cables come out of the cpu block for some unimaginable reason.

[WoodenMarker]

1 hour ago, Jon-Slow said:

I am willing to buy new stuff. For the CPU cooler I did a lot of research and Scythe Mugen 5 rev.b  seemed to on the same level of performance as the Noctua NH-D15, only slightly cheaper. I don't know if there would have been much difference unless NH-D15 is noticeably better. As for the Thermal paste, I am looking into it but I'm not sure how much of an improvement that is. currently, was looking into buying a 'thermal grizzly conductonaut liquid metal' but I'm not sure how much better it would be compared to the 'thermal grizzly kryonaut' or my current Scythe Mugen stock thermal solution or my 2 years old 'Cooler Master E2 IC Essential'.

I am also willing to look into liquid or water cooling but I'm not sure if they can give me any advantages since most reviews suggests I should stick with the Air cooling.

Yes, the NH-D15 is easily better. Saying the Mugen 5 Rev. B performs the same as the NH-D15 is like saying the Hyper 212 Evo performs the same as the Mugen 5 Rev. B. 

What settings have you changed form default and how did you arrive at 1.35v to begin with?

What do you mean by temps getting too high? Are they going higher than 85c? 85c isn't very low but it's perfectly safe. 

[JonS77]

3 hours ago, WoodenMarker said:

Yes, the NH-D15 is easily better. Saying the Mugen 5 Rev. B performs the same as the NH-D15 is like saying the Hyper 212 Evo performs the same as the Mugen 5 Rev. B. 

After seeing reviews like this one it was difficult for me to justify paying extra for the NH-D15 even though I was almost confident the NH-d15 is what I'll buy. So if I have made a mistake and the reviews wouldn't thoroughly cover how much better the NH-D15 would have been I'm willing to take the L and make another purchase. I've had a difficult time seeing what the NH-D15 would do in comparison to this one considering its price point. For example, in the tweaktown review the NH-D15 was 1.75 degrees cooler under overclock load and its noise was similar to Mugen 5. So I got a little confused and purchased this one instead. I'm still open to finding out a test or a benchmark that could show me where I've made the mistake and make me redo my purchase.

 

3 hours ago, WoodenMarker said:

What settings have you changed form default and how did you arrive at 1.35v to begin with?

Turing on the XMP profile sets the voltage at 1.35v and turning it off lowers it to 1.243. It also adjust the memory frequency to 3000MHz from its stock speed 2133. This results in a higher work load on my CPU and so higher temps.

 

3 hours ago, WoodenMarker said:

What do you mean by temps getting too high? Are they going higher than 85c? 85c isn't very low but it's perfectly safe. 

idle temps   = 35°

normal load = 78°

Stressing the CPU with CPUz after a few minutes = 81°

AC Odyssey = 78°  (this is the benchmark, but it could go as high as 80° while playing for an hour)

AC Odyssey (5Ghz) = 85°- 89° (and somehow the performance is worse, the CPU is obviously unstable at this point)

Cinebench = 80°

Cinebench (5GHz) = 100° 

 

The chassis ambient temp could fluctuate between 15-28° under different circumstances.

 

If I could find similar results with the NH-D15 it could be very helpful.

 

So considering all this, I don't really have a CPU that could run at 5GHz and even the 4.7GHz seems problematic. I really feel like I've made a huge mistake going for the 9900k.

[WoodenMarker]

10 hours ago, Jon-Slow said:

After seeing reviews like this one it was difficult for me to justify paying extra for the NH-D15 even though I was almost confident the NH-d15 is what I'll buy. So if I have made a mistake and the reviews wouldn't thoroughly cover how much better the NH-D15 would have been I'm willing to take the L and make another purchase. I've had a difficult time seeing what the NH-D15 would do in comparison to this one considering its price point. For example, in the tweaktown review the NH-D15 was 1.75 degrees cooler under overclock load and its noise was similar to Mugen 5. So I got a little confused and purchased this one instead. I'm still open to finding out a test or a benchmark that could show me where I've made the mistake and make me redo my purchase.

 

Turing on the XMP profile sets the voltage at 1.35v and turning it off lowers it to 1.243. It also adjust the memory frequency to 3000MHz from its stock speed 2133. This results in a higher work load on my CPU and so higher temps.

 

If I could find similar results with the NH-D15 it could be very helpful.

So considering all this, I don't really have a CPU that could run at 5GHz and even the 4.7GHz seems problematic. I really feel like I've made a huge mistake going for the 9900k.

That's one review out of many. Most show the NH-D15 clearly ahead and this holds more true as the loads get heavier. 70c is fairly moderate for a load temp and doesn't really let coolers stretch their legs. The higher the temperature, the greater the delta in temps is when comparing the same coolers. 

At low loads, differences aren't clear -- similarly to how all coolers idle around the same temp. As loads and temps increase, the difference between each cooler becomes clear until the weakest cooler fails too keeps temps in check. 

Here are some more examples between the two: 

https://www.realhardwarereviews.com/scythe-mugen-rev-b-review/6/

https://www.hwcooling.net/en/scythe-fuma-rev-b-surprise-with-a-big-question-mark/4/

https://www.hardwareluxx.de/index.php/artikel/hardware/kuehlung/41102-scythe-mugen-5-im-test.html?start=5

https://www.conseil-config.com/2017/test-scythe-mugen-5/4/#les-temperatures

There are many other reviews that can be used to assess how the coolers perform. If you want to see how coolers compare in performance, here are the results:

 

You should be able to load the xmp profile for the ram only. Manually oc the cpu instead. 

Are you following any oc guides?

 

There are plenty of other variables that affect cooling including ambient temp. It's difficult to draw any strong conclusions from looking at other configurations. 

You can see examples of systems that use particular components on pcpartpicker. For example, the NH-D15 with oc'ed 9900k's: https://pcpartpicker.com/builds/by_part/jHZFf7#h=209&C=1

[JonS77]

17 hours ago, WoodenMarker said:

There are plenty of other variables that affect cooling including ambient temp. It's difficult to draw any strong conclusions from looking at other configurations. 

You can see examples of systems that use particular components on pcpartpicker. For example, the NH-D15 with oc'ed 9900k's: https://pcpartpicker.com/builds/by_part/jHZFf7#h=209&C=1

Alright, I guess I will make the payment for an NH-D15 now.

 

Also, my ambient is often normal room temperature but when I want to test I turn on the AC and make the room cold and still get not much difference in my temps.

[WoodenMarker]

6 hours ago, Jon-Slow said:

my ambient is often normal room temperature but when I want to test I turn on the AC and make the room cold and still get not much difference in my temps.

So by normal room temp, you mean 21c?

Component temps generally scale directly with ambient temps. A 5c rise in ambient will usually mean a 5c rise in component temps.