Status
Not open for further replies.

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
I bought the H115 Platinum since the stock cooler was loud as hell. I am using the Thermaltake Core P3 and my original temps with the PrismCooler were 45-55 idle and hitting 70+ at Cinebench R15.

Now that i got the H115, i installed it and saw it already came with thermal compound applied. I cleaned old one with ArctiClean and installed. I do still have my Artic Silver 5 but i didn't bother since this had some already pre-applied.

https://ibb.co/dJsMnzX

My highest score was 3127 and after installing iCue my score dropped to 3053. My temps on idle are 38-48 and in cinebench easily hitting 77C. This doesn't seem right imo, i uninstalled the iCue to see any difference but no avail. Please correct me if the temps with this cooler shouldn't even hit 70C?

Also i didnt touch the fan profile in the iCue, it was default to "silent" for fans and "balanced" for the pump. Putting them both on extreme prevented CPU going above 72C but then you have a full jet engine sound.

Whats going on here? iCue broke cinebench or is my cpu suffering? Should i remove the cooler and add my own thermal compound?

P.s
I ran the R20 and the score is 7057, pretty much identical to what others are getting, but still temp hit 76C without the iCue since i uninstalled it.

https://ibb.co/9v9vDvz
 
Last edited:
Solution
If you're going to do that I'd just get another set of the same Maglev fans you already have and put them on the other side of the radiator pulling with the first set pushing.
If those are full load stress test temperatures, then yes, they are normal. They might even be normal for peak gaming temperatures. Much depends not on the cooler or paste, but what your ambient temperatures are because any cooler you use will not be capable of "this temperature" but instead will be resulting in "this temperature OVER ambient".

Also, your fan profile configuration for the cooler will play a role as well and you may need to play around with your fan curve to tweak your thermal results a bit to your own liking.

For that CPU, anything under 80°C is fine for a peak load temperature while running 100% on all cores. If you are getting at or above 80°C under any conditions, then something is wrong. Especially if you are not overclocking. If you have PBO enabled, that is obviously going to increase thermals under most conditions. If you have a high ambient temperature in the room where your PC is, then you are likely NOT going to get the same thermal result as somebody who lives in a cooler region would.
 
CPU_FAN and CPU_OPT are the same for most boards. Some boards REQUIRE that there be a fan connected to CPU_FAN before CPU_OPT can be used. I would recommend that use CPU_FAN rather than CPU_OPT and that you adjust the fan curve or preset so that the CPU_OPT header running the pump is at full speed 100% at all times while turned on.

Some boards have a dedicated PUMP header or AIO_PUMP header, and when those are present it is wise to use that.

Even if you add two more fans for a push pull configuration, you would want to turn them the other way. I'll explain because I know you are not understanding the different between sucking air IN through those restrictions and trying to blow air OUT through them.

Pulling air in through a restriction works better than trying to push out through it. The best way I can explain this is with the string comparison.

Example. Get a piece of string about three feet long. Now, while holding one end, leave the other end on the ground and pull it. Works fine right? No problem. You can pull the string all day and it will simply follow wherever you go.

Now, take the same string and try to PUSH it. LOL. Yeah, doesn't really work does it?

While that is not a direct comparison to the aerodynamics of pushing or pulling through an orifice, it does give you an idea of how pushing and pulling can have different effects on the same media.

No matter how many fans you use, and no matter what side of the radiator they are on, you want them either pushing away from the orifice/restriction THROUGH the radiator and away from that case framework OR pulling through the radiator and blowing off into the open air. Plus if you go with a push pull configuration then a whole other set of factors come into play such as having fans on the pull side that have a higher CFM capability than the fans on the push side AND having fans on the push side that have a very good static pressure rating.

Honestly, if it's working now and the aesthetics are that important to you, then leave it like it is. If you want to get the most performance out of your cooler, then either leave the fans where they are but flip them so that they are blowing away from the radiator, or put them on the other side of the radiator blowing THROUGH it.
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
My room temp is 23 C and no idea if PBO is enabled. I didn't touch anything in bios other than XMP for ram when i first installed. But my hope was with a massive cooler the max temps under load for a min wont exceed 70C. Because its not that i am running some stress test for 2 hours, just cinebench that takes merely 45 seconds or something to finish.

Sorry i may not be too familiar with temps, correct me if my hopes were too high.
 
70 really isn't bad if that is the maximum temp under stress. Even 80°C on that CPU would be acceptable.

I think I would probably take it back apart, remove all of the pre-applied thermal compound and completely clean the bottom of the water block and the top of the CPU heat spreader and use a regular good thermal compound like your Arctic Silver 5. I would recommend applying about as much as half of a frozen or snow pea right in the center of the CPU lid. Be sure that your backplate is fully seated/secure and that you completely fasten down the water block to the backplate mounting hardware but don't overtighten it. Snug but not overtightened is the key. You will get a "feel" for when it is tight enough. It does not need to be "cranked down", just tight.
 

jon96789

Reputable
Aug 17, 2019
414
49
4,740
I also have a AMD Ryzen 9 3900X with a Corsair H115i RGB Platinum cooler. Under load when encoding or rendering videos, the CPU temp will hover around 90 degrees. With the stock AMD Wraith Prism cooler, i was hitting 95+ degrees. With the stock thermal paste, it would drop to about 39 degrees.

I replaced the stock paste with Thermal Grizzly's Kryonaut to see if that would help matters... Sad to say, my idle temp actually went up to about 43 degrees but peak temps remained the same. I would not replace your thermal paste unless necessary. I also ordered Noctua's NT-H2 paste and will see if that makes any difference.
 
I'd say you have a problem with your cooler or it's incorrectly mounted. Or the paste job was done improperly. Maybe an air bubble or weak pump. Could be a poorly configured airflow arrangement depending on the case and how the rest of the cooling system is setup but that is not a typical expectatation for a 280mm Corsair AIO cooler even with a 3900x. Possibly there are issues with the core voltage as well. Something isn't right, because that cooler is far more capable than what your results seem to suggest AND that CPU isn't THAT damn hot with only a 105w TDP unless you're overclocking or have PBO enabled.

Unfortunately, I don't have one of those setups to confirm or deny, but those results are not in line with what I see others achieving with similar configurations AND your temps are beyond the maximum recommended thermal specs so I would definitely not continue to run it that way until you sort out your thermal issues.
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
Just a heads up, i did enable the PBO and my score jumped by 120+ points in cinebench. My temps hit max 81.75C during the R20 test and idle at 40-50C. Without the PBO as i mentioned earlier, i was hitting MAX 72-76C depending if I had my AC on/off.

So the question again is, 82C safe? or should i dial back the PBO? I don't want to hurt the cpu in any way, what is your sugegstion? Im happy with -100 score in cinebench lol.. and not sure if all this does give any edge in gaming.

Thanks
 
82 is NOT safe. Idle at 40-50 degrees Celcius is also NOT normal. There is a problem with a loose backplate, or bad paste job, or something. Even if you were peaking above the desired temperature while under a load, your idle temps if everything were correctly installed and functioning normally should not typically be that high. I'd double check everything including the connections and the software configuration and fan curves.

How do you have this installed, as an intake orientation, or an exhaust orientation, and WHERE on the case is the radiator located?

Regarding thermals on Ryzen, everybody needs to read THIS entire article, because even 80°C may be too hot according to AMD.

https://www.tomshardware.com/news/amd-ryzen-3000-boost-clock-controversy-intel-attack,40231.html
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
82 is NOT safe. Idle at 40-50 degrees Celcius is also NOT normal. There is a problem with a loose backplate, or bad paste job, or something. Even if you were peaking above the desired temperature while under a load, your idle temps if everything were correctly installed and functioning normally should not typically be that high. I'd double check everything including the connections and the software configuration and fan curves.

How do you have this installed, as an intake orientation, or an exhaust orientation, and WHERE on the case is the radiator located?

Regarding thermals on Ryzen, everybody needs to read THIS entire article, because even 80°C may be too hot according to AMD.

https://www.tomshardware.com/news/amd-ryzen-3000-boost-clock-controversy-intel-attack,40231.html

I have an open air case, and mounted like this:
https://ibb.co/Q9VM47t

As i said i am using the thermal compound that came with the AIO. I can take it off and apply Artic Silver 5. My fan profile is on "Balanced". I thought u said anything below 80C is normal. As i said my max is hitting 82C under benchmark with POB on. Without it was 70ish something under stress.

My idle temp at the moment writing this with chrome open is 37C - 44C. Room temperature is 22C.
 

jon96789

Reputable
Aug 17, 2019
414
49
4,740
I believe people do not realize that the 3900X runs a lot hotter than previous chips... Same as the Intel i9 9900K... My prebuilt system ran 40 degrees at idle and 95 degrees under load with the AMD Prism cooler. With the Corsair H115i it peaked at 85 degrees...
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
I believe people do not realize that the 3900X runs a lot hotter than previous chips... Same as the Intel i9 9900K... My prebuilt system ran 40 degrees at idle and 95 degrees under load with the AMD Prism cooler. With the Corsair H115i it peaked at 85 degrees...

So we are identical! With POB i have 82C and under benchtest and without the POB low 70's.
 
If you read the article I linked to in it's entirety, you will see that it is not what "I" said, it is what AMD said. Then they changed the thermal limit to 70°C because, apparently, they were seeing indications of longevity issues at higher temps. I can't control what the CPU manufacturer decides from week to week is the safe temperature and what temperature it should throttle at. Totally not my words, but theirs. If Intel suddenly decides that for their CPUs a different threshold is appropriate, then we'll be having this same discussion about theirs as well, but I doubt that would happen since 80°C has been the recommended thermal limit for many years on all Core-i processors.

I realize that the 3900x has more cores and therefore runs hotter. That does not change the fact that all of the Zen2 CPUs have the exact same thermal limit, and just because it has more cores and runs hotter does mean it is OK for it to be run hotter. It's not. With the latest BIOS versions installed those chips SHOULD be throttling at anything over 70°C based on the information in that article. Obviously, they are not, OR neither of you have the latest BIOS versions installed, OR, your motherboard manufacturer has declined to implement the latest microcode that changed the throttle limit from 80 to 70. IDK, I can't answer that. All I CAN say is that if you are seeing temperatures in the 90's, there IS a problem. That is an indisputable fact no matter what BIOS version or motherboard you have.

If you wouldn't mind, post a picture of the opposite side of your case, and also please explain why your radiator and fan configuration are installed in a manner that has the radiator fans trying to not only blow through the restriction of the radiator, but also through the restriction of whatever is on the other side of the radiator. Technically, depending on what is on the other side of the radiator, the fans should be in between the case and the radiator, blowing through the radiator and into the open space inside your case. I assume you arranged them the way you did for aesthetic reasons but I'd almost guarantee that practically half the potential airflow of those fans is actually making it through there. The fans should either be flipped over so that the are sucking through the radiator OR moved to the other side of the radiator blowing through it.
 
Tested over AGESA 1003ABB

From what i tested on my cheap Ryzen 5 3600, I can get 4.2GHz on single core (with PB enable only ) without an issue (either with stock cooler or the tower cooler I got).
After swapping the stock cooler to a tower cooler Gammaxx GTE (nothing as fancier as bequiet or noctua), my all core frecuency went up while running Cinebench R20 to around 4075 to 4125 GHz, and after 30mins of consecutive Cinebench runs max temp was 72.8°C (used hwinfo to get the readings, around 18~22°C ambient temp).

Once again, in my case the only thing I got with PBO and Auto overclock enable was a higher vcore voltage, a higher temp and no performance gain. So almost since day one I just keep all PBO and AutoOC options in BIOS OFF.

As for what the article said, it will be good if we finally find out whats happening for a group of AMD users.

But once again, a 2700ish total entries from Der8auer's survey is not bad, but is not a huge amount of results to get a "real" scenario of whats going on in every single chip segment.
If I remember correctly there were only around 600 results for the Ryzen 3900X, and theres no way to make sure every single one of those was tested correctly.

mean, Im not AMD nor Intel fan boy, Im just saying that perhaps not everyone that uploaded their result knew how to properly install Windows and Drivers, or how to properly setup the options in BIOS, maybe some of them didn't even reinstalll windows after swaping an intel mobo+cpu to a new Ryzen (we saw at least 1 post in this same forum about it), thye could have unknown software running on the background eating resources, imposible to know.

Anyways, I hope this turn out well for we "the users" who put our fate and money on thier products.
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
If you read the article I linked to in it's entirety, you will see that it is not what "I" said, it is what AMD said. Then they changed the thermal limit to 70°C because, apparently, they were seeing indications of longevity issues at higher temps. I can't control what the CPU manufacturer decides from week to week is the safe temperature and what temperature it should throttle at. Totally not my words, but theirs. If Intel suddenly decides that for their CPUs a different threshold is appropriate, then we'll be having this same discussion about theirs as well, but I doubt that would happen since 80°C has been the recommended thermal limit for many years on all Core-i processors.

I realize that the 3900x has more cores and therefore runs hotter. That does not change the fact that all of the Zen2 CPUs have the exact same thermal limit, and just because it has more cores and runs hotter does mean it is OK for it to be run hotter. It's not. With the latest BIOS versions installed those chips SHOULD be throttling at anything over 70°C based on the information in that article. Obviously, they are not, OR neither of you have the latest BIOS versions installed, OR, your motherboard manufacturer has declined to implement the latest microcode that changed the throttle limit from 80 to 70. IDK, I can't answer that. All I CAN say is that if you are seeing temperatures in the 90's, there IS a problem. That is an indisputable fact no matter what BIOS version or motherboard you have.

If you wouldn't mind, post a picture of the opposite side of your case, and also please explain why your radiator and fan configuration are installed in a manner that has the radiator fans trying to not only blow through the restriction of the radiator, but also through the restriction of whatever is on the other side of the radiator. Technically, depending on what is on the other side of the radiator, the fans should be in between the case and the radiator, blowing through the radiator and into the open space inside your case. I assume you arranged them the way you did for aesthetic reasons but I'd almost guarantee that practically half the potential airflow of those fans is actually making it through there. The fans should either be flipped over so that the are sucking through the radiator OR moved to the other side of the radiator blowing through it.


Here you go sir:
https://ibb.co/ySWph8s
https://ibb.co/QdkX2Xk

As i mentioned before i am 70C without POB but with it enable hitting 82C :(. Have at look at the case, if my temps are uniquely that bad, i will take it apart and re-apply thermal compound.
 
You need to change that around. For one, that's an incredibly lame design for liquid cooling because no matter which direction those fan blow there is going to be a restiction due to the design of that grill on the case but it won't be AS bad if the fans are sucking air IN through that grill as it is with them trying to BLOW through it.

My advice would be to take the radiator off and move the fans to the other side of the radiator with the open side of the fans facing the grill on the case so that they are blowing IN through the radiator. Not surprising that Thermaltake has chosen to completely bork their own case design up by doing dumb crap like that.

There is no way your cooling system can perform well with the way it is mounted right now. No. Way. Possible. You are simply shooting yourself in the foot.
 

zikmir

Distinguished
Jun 29, 2013
50
8
18,545
I removed the back cover to have complete open air flow like this:

https://ibb.co/xMQJyv0

Made ZERO difference in temp, for the exact same 82C under Cinebench with POB Enabled. For now i have disabled POB and i dont hit more than 71C in POB at cost of 50 points.

So what are your thoughts? I dont think that grill restricts anything, since i got exact same temps and found no forums people stating difference or bad cooling.
 
You think what you want to think. I really could care less, but I'm telling your FOR A FACT that any fan, high static pressure or not, that has to blow THROUGH those kinds of restrictions, and I'd bet that YOUR fans, are NOT high static pressure fans to any serious degree, is going to suffer from it.

There's a reason why people mod cases to get rid of the solid front panels to improve airflow both with air cooled and water cooled systems. Removing that back panel isn't going to do that much. The system is still seriously restricted by having to blow not only through the restriction of the radiator itself, but also through the restriction of only a 2/3 open and seriously narrowed opening. You need to move those fans, WITH the panel removed like it is now, to the other side of the radiator sucking air IN through those openings and out through the radiator into the open air. I don't get, 100%, why it's hard to understand.

These are basic fundamentals and you have them completely reversed from how they should be, so, if the looks from the fans are more important to you than the actual cooling, then I guess this thread has run it's course and there is no reason to continue it because you are not going to EVER improve your cooling performance with the radiator and fans configured the way they are now. You're not dealing with some first time builder noob with no knowledge of how restrictions affect cooling. Jesus. Why do you people even come here and ask questions if you don't want to listen to the advice that's offered. It might take you what, ten minutes to reconfigure those fans and see if it makes a difference at all, which it definitely should.

Especially if you haven't also failed to configure the pump correctly. You do have the pump connected to the CPU_FAN header or a dedicated pump header on the motherboard and turned to full time 100% speed operation in the BIOS yes? Because if you DON'T, then that's another thing not done right.
 

Karadjgne

Titan
Ambassador
Ok. To start with there needs to be an understanding of what PBO is and what it does. Ryzens/TR are what's considered a 'smart' cpu in that it can and does use thousands of sensors to monitor its activity and attributes. But it also has built in limits. PBO understands those limits, and if available will mess with 3 things. SoC, temps, VRM's. The PBO will test those sensors over 1000x a second and if it finds room to grow in any or all of those three, it will push the cpu faster on more/all cores until limits are reached.

That means voltages will go up, temps will go up, all in an effort to raise performance.

Meaning that with PBO active, that cpu is always going to see maximum available temps and high voltages every time you push the cpu with high loads like stress tests as it tries to get that extra few Hz per core.

Changing paste won't help, changing coolers, fans, location won't help. If there's an extra 10°C of room, the cpu will use it and you'll be at/over 80°C every time. If voltages are lowered in OC, and there's space left, same deal, PBO will crank them up to get more speed per core.

The only way you'll drop temps is by disable PBO, remove the temptation of the cpu to push its limits in favor of limits you set.

Having decent Cinebench scores is nice, but unless that's the only reason you do this, they don't have any affect on real world scenarios. You'd be far better off with a non-PBO static OC running mid 70's than a PBO pushing limits to shave a few seconds on a long render at over 80°C.

As to that extra 10°C of room, that's going to be up to the cooler and it's location. The P3 is somewhat limited for 360mm AIO's, they only go in 2 places. The standard placement is on the right side as shown in half a hundred pictures. This does present issues. There's basically an 80mm slot behind a 120mm wide radiator. This can and does put a crimp on airflow pushed into it as the sides block air coming out the edges of the rad. This lowers available flow, creates backpressure which lowers flow and makes more work for the fans. Pull on the other hand is far more successful in moving air from a restricted space to open space. Turn the fans facing the rad, and you'll get better flow through the rad, better capacity and probably better temps in non-PBO.

If you want to see the RGB on those pretty fans, get some spacers about 1"-1.5" and put the rad off that backplate and let it breathe.

The P3 was designed more for looks than performance
 
Last edited:
  • Like
Reactions: RodroX
And that is what I am saying. Either flip the fans over so they are pulling through the radiator from a restriction into open air, or move them to the other side of the radiator so they are still pulling through the restriction (Not trying to BLOW through a restriction, which poses far more problems) but are blowing through the radiator into open air.

If you had fans with insane static pressure like the Noctua iPPC industrial 3000rpm fans, or the EVGA 400-HY-FX12-KR fans with 4mm H20 static pressure, you MIGHT get away with it being the way it is now, but even then you would still be drastically reducing the cooling performance that you COULD be seeing. With the fans you have now, you are simply blowing about 30% of the air nowhere because it cannot overcome that level of static pressure resistance. Turned the other way, there would be almost no resistance at all.
 
Status
Not open for further replies.