Crazy-high Ryzen 3900X temperatures (95c immediately)

Maybe they will make the heat spreaders out of diamonds as they conduct heat better then copper. But density and surface area are becoming a big problem now.
 
  • Like
Reactions: N4CR
like this
What worked in the past may not work good today - Ryzen 3 is a multi chipped CPU, each chiplet area on the heatspreader will need paste. Pee size from my experience rarely gave full coverage of the heat spreader but for older design CPUs it would cover the die area. Not so with Ryzen 3's chiplets. This was also the case with Tread Ripper where pea size for the rather large chip failed. I use a thermal pad which covers the full chip, my idle temps are in the 30's, CineBench it goes to around 70c. A good paste job would probably drop me 1c-2c.

Now Ryzen 3 boost clock is very sensitive to temperature and thus performance really can be massively affected by how good you are cooling all of the chiplets.
 
My H110 failed just before the 3 year period was up and I had it replaced by Corsair. Still using the replacement that is now almost 3 years old in its own right. Symptoms: My 4770K would shut down due to overheating before it got out of the BIOS.
 
Krenum said:
Get a Noctua NH-DH15 and chunk the AIO garbage.
Click to expand...

some people including me prefer the small footprint of an aio rather than looking at a boat anchor hanging off a mainboard. dont get me started on those ugly brown fans either.
 
Riccochet said:
This is why Der8auer suggested spreading paste in to a thin layer across the entire IHS. Kryonaut comes with a little spatula just for that. It's how I've been doing it for probably 10 years now. You actually use less TIM since you only need a very thin layer. I usually make it damn near transparent.
Click to expand...

I used to do this until I watched a video where they used a piece of glass so you could see how the goop spreads. Any coat makes air bubbles and its almost impossible to avoid. Doesn't do much temp wise, but it bothers me. So now I do big glob in center. Not tiling a floor.
 
Krenum said:
Obviously you don't know how the mounting bracket works. And also, even if the fans fail, the heatsink is large enough to cool your CPU passively. I agree, its quite fool proof.
Click to expand...

I've had them fail before, I know very well how they work :) thanks though. I might not have mounted a DH15, but I've mounted some heavy HS over the years, including Snapping AMD board mounting brackets.
 
Krenum said:
Obviously you don't know how the mounting bracket works. And also, even if the fans fail, the heatsink is large enough to cool your CPU passively. I agree, its quite fool proof.
Click to expand...

The heat sink will allow you to go a bit longer before throttling than the water block alone would, but either way you're replacing shit.

There's nothing wrong with air cooling, but to claim its 100% is retarded.
 
Digital Viper-X- said:
I've had them fail before, I know very well how they work :) thanks though. I might not have mounted a DH15, but I've mounted some heavy HS over the years, including Snapping AMD board mounting brackets.
Click to expand...

First, sorry for your loss, you should read instructions and use proper tools. Second, if you haven't used it, then you don't know what your'e talking about and you don't have an argument.
 
RamonGTP said:
The heat sink will allow you to go a bit longer before throttling than the water block alone would, but either way you're replacing shit.

There's nothing wrong with air cooling, but to claim its 100% is retarded.
Click to expand...

Why would you have to replace something? Only a moron would continue to run their PC under load knowing that the fans aren't working.
 
Krenum said:
Why would you have to replace something? Only a moron would continue to run their PC under load knowing that the fans aren't working.
Click to expand...

If your fan isn't working you have to replace the fan smart guy... Jesus.
If a pump fails you replace the pump, or in the case of AIO, the whole unit.

Both solutions work, both have their pros, both have their cons, and neither is 100%
 
RamonGTP said:
If your fan isn't working you have to replace the fan smart guy... Jesus.
If a pump fails you replace the pump, or in the case of AIO, the whole unit.

Both solutions work, both have their pros, both have their cons, and neither is 100%
Click to expand...

Oh, well yeah. I thought you were implying that you'd have to replace a major component. I think we can all agree that a nickle/copper block is more effective upon failure of an air supply than a plastic pump.

But really though, how often do fans fail? From my experience in 25 years of building PC's I can count on my one hand that a fan has completely died.
 
RamonGTP said:
If your fan isn't working you have to replace the fan smart guy... Jesus.
If a pump fails you replace the pump, or in the case of AIO, the whole unit.

Both solutions work, both have their pros, both have their cons, and neither is 100%
Click to expand...

You can still use your machine without fans on the CPU heatsink. Probably for awhile

Meanwhile if your pump goes it'll be seconds before you start hearing the BIOS alarm
 
Krenum said:
First, sorry for your loss, you should read instructions and use proper tools. Second, if you haven't used it, then you don't know what your'e talking about and you don't have an argument.
Click to expand...

Going by your logic, have you used the AIO in question? or tested every AIO long term to verify that they are less reliable than HSF units ? Asking for a friend
I didn't do anything wrong, :) it was fine for over a year and one day temps went crazy, the bracket snapped, shit, I guess sometimes things break don't they?
 
blackmomba said:
You can still use your machine without fans on the CPU heatsink. Probably for awhile

Meanwhile if your pump goes it'll be seconds before you start hearing the BIOS alarm
Click to expand...

Hmm, not sure about this one, but it is easy to test, maybe Krenum here can run the test for us, unplug the fans, see how long it will last under normal use, then see how long it will last under gaming use (make sure to dump the heat between runs)
 
Digital Viper-X- said:
Going by your logic, have you used the AIO in question? or tested every AIO long term to verify that they are less reliable than HSF units ? Asking for a friend
I didn't do anything wrong, :) it was fine for over a year and one day temps went crazy, the bracket snapped, shit, I guess sometimes things break don't they?
Click to expand...

Was the bracket plastic? Plastic doesn't do well with heat, it becomes brittle, if it was then Im not surprised it snapped.

I've used many AIO's, they all have the same design. And they've all failed or leaked or smelled like a fishtank. They are garbage.
 
Krenum said:
Was the bracket plastic? Plastic doesn't do well with heat, it becomes brittle, if it was then Im not surprised it snapped.

I've used many AIO's, they all have the same design. And they've all failed or leaked or smelled like a fishtank. They are garbage.
Click to expand...

Yes they are plastic, I've used 3 AIOs, and am still using 2 of them today, none of the 3 have failed yet(well maybe the 3rd failed, I don't have it anymore so I can't comment :) ) but my other 2 , which are 24 x 7 use ones are still going. (going on 4 years now on the first, and 3 on the second)



I'm sure we've all used these before.
 
Digital Viper-X- said:
Hmm, not sure about this one, but it is easy to test, maybe Krenum here can run the test for us, unplug the fans, see how long it will last under normal use, then see how long it will last under gaming use (make sure to dump the heat between runs)
Click to expand...

Hmm, if I get time over the weekend I may try it, that's a good idea.
 
Krenum said:
Why would you have to replace something? Only a moron would continue to run their PC under load knowing that the fans aren't working.
Click to expand...

Not much different from a moron that runs with a broken AIO as well.


Krenum said:
Was the bracket plastic? Plastic doesn't do well with heat, it becomes brittle, if it was then Im not surprised it snapped.

I've used many AIO's, they all have the same design. And they've all failed or leaked or smelled like a fishtank. They are garbage.
Click to expand...

All failed or leaked? Fascinating. What is your sample size?

I have 5 AIO's in my office, some as old as 2012 that still work.
 
Digital Viper-X- said:
Yes they are plastic, I've used 3 AIOs, and am still using 2 of them today, none of the 3 have failed yet(well maybe the 3rd failed, I don't have it anymore so I can't comment :) ) but my other 2 , which are 24 x 7 use ones are still going. (going on 4 years now on the first, and 3 on the second)



I'm sure we've all used these before.
Click to expand...


Aye, the NH-D15 is all metal with tension screws. Once its locked it will never loosen unless you torque it.
 
thebufenator said:
Not much different from a moron that runs with a broken AIO as well.




All failed or leaked? Fascinating. What is your sample size?

I have 5 AIO's in my office, some as old as 2012 that still work.
Click to expand...

60's & 120's x2 (240). You are a brave man. I wouldn't trust them. Or I guess I've just had bad luck with them.
 
rhexis said:
some people including me prefer the small footprint of an aio rather than looking at a boat anchor hanging off a mainboard. dont get me started on those ugly brown fans either.
Click to expand...

And some prefer lots of bare shiny fins versus the poor CPU being stuck in ICU with life-support hoses sticking out at weird angles, strange medical dialysis pumps spurting bad koolaid and the imminent threat of rgb blindness. :D
 
These density/temperature realities kinda make you wonder if the IHS is gonna need to be re-approached. Neither AMD or Intel has changed the design of them in 15~ years, seemingly. You've gotta wonder how much of a hindrance they are to controlling temperatures at this point.

And isn't most of the reason we've never seen the IHS approach make it to GPUs... density/temperature?

Has anybody done delid/bare-die contact cooling tests with Ryzen 3000 yet?
 
Centauri said:
And isn't most of the reason we've never seen the IHS approach make it to GPUs... density/temperature?
Click to expand...

Good point. Haven't really thought about that since you mentioned it, probably because its been the norm for so long. I'd imagine here in the next 5-10 years they'll have to do something to address it.
 
hititnquitit said:
Ive only seen one dude delid his 3600 and he only gained 4c in doing it.
Click to expand...

Yeah but everybody who delids is basically just doing it to test with liquid metal between the die and the IHS instead. I'm talking about deliding and ditching the IHS completely. Heatsink-to-die contact like it was back in the Athlon XP days.

Not something I'm willing to try myself at all, haha, and it would certainly require some heatsink hardware mount modification to adjust for the height difference, but I'd be verrrryyy curious to see the results of a test like that.
 
Krenum said:
I'd rather have a 100% fail proof application.
Click to expand...

Neither are fail proof has the large chunk of metal can warp and damage motherboards over time due to its weight (assuming you’re looking for one that can compete with an aio).
 
VIC-20 said:
I used to do this until I watched a video where they used a piece of glass so you could see how the goop spreads. Any coat makes air bubbles and its almost impossible to avoid. Doesn't do much temp wise, but it bothers me. So now I do big glob in center. Not tiling a floor.
Click to expand...
Once the heatsink is fully secured (high pressure) and the CPU is loaded (high heat), that air moves out.

Those videos with glass plates are only good for a general idea of how each application method spreads for coverage. That is all.
 
Gideon said:
Maybe they will make the heat spreaders out of diamonds as they conduct heat better then copper. But density and surface area are becoming a big problem now.
Click to expand...
Price is the issue. Maybe a decade From now. The low grade polycrystalline stuff isn't super conductive. Monocrystal is. I have some optical grade at 2200 W/mK I pumped with nearly sun power density levels via laser (current visible wr) and it took it like a champ.thermal gradients are exceeding even theoretical expectations with that material.
 
Digital Viper-X- said:
Hmm, not sure about this one, but it is easy to test, maybe Krenum here can run the test for us, unplug the fans, see how long it will last under normal use, then see how long it will last under gaming use (make sure to dump the heat between runs)
Click to expand...
I'll do it. But my hsf is designed to run passive ;)
 
hititnquitit said:
Ive only seen one dude delid his 3600 and he only gained 4c in doing it.
Click to expand...
It won't change much. People forget limitation is heatsink base and conductivity of silicon. We are rapidly approaching silicon power density vs conductivity. ( extracting that heat density) trouble. Zen2 has density like an oc'd older process.

I would go exotic spreader or a vapour chamber or pipe design. Then a 500w water cooled pelt.
Thermal gradient would be insane and I think Will help a lot. Chilled water is final step.. Be a nice daily able system.
 
Centauri said:
Yeah but everybody who delids is basically just doing it to test with liquid metal between the die and the IHS instead. I'm talking about deliding and ditching the IHS completely. Heatsink-to-die contact like it was back in the Athlon XP days.

Not something I'm willing to try myself at all, haha, and it would certainly require some heatsink hardware mount modification to adjust for the height difference, but I'd be verrrryyy curious to see the results of a test like that.
Click to expand...

I did this with spectacular results on my golden 3770k 5Ghz AC OC (everything stable) and 5.2Ghz (every game I ever played stable) with a 16GB kit of ddr3 2400 running at 2600 cas 11...

It worked really well, I used the stock XSPC Raystorm retention bracket and some very thin stainless steel washers to offset the height loss of the IHS...I used CL LM pro and my temps were never above 62c even running P95 with an AVX load.

The only issue with running direct die cooling is that if downward pressure on the CPU...I basically used a trial and error approach with mine...

I set the CPU, LM applied, and put the GPU block on top...I then slowly tightened it and kept attempting to boot every 2 turns...

It will boot at first but only have single channel memory so you have to ensure that it has just enough pressure where all the pins make contact but not so much that you warp the PCB substrate abd potentially kill the CPU..

This was much less of an issue with the XP era since they used a PGA package type where the pins were on the CPU vs a LGA type where the pins are in the socket..The fact that AMD went back to a LGA type for AM4 means that it would be much easier for the mainstream crowd to deal with it. They just need to include a laser cut shim so people do not crack their die.


Since most gamers these days use AIOs vs a 1.5kg hunk of metal I think there would be much less risk to using a bare die setup.
 
Krenum said:
Click to expand...

You can cling to your horse and buggy instead of following the rest of the world into modern times but a proper loop is the best you can get before moving to exotic cooling methods...

If an AIO raped a family member, I apologize for their lack of decency but plenty of people use them with great results and do not suffer from a high level of failures. If that were true then you would not see 50+ companies with a product in the channel.
 
You must log in or register to reply here.
Back
Top