[UPDATED] Der8auer States Some X299 Low End boards Have Bad VRM Coolers

[Hunter259]

Just now, cj09beira said:

Vrms heat production scales much more with amps than with volts

That is true but there is still heat added onto the board that on X299 is on the CPU. The ThreadRipper board is going to need a better cooling solution.

[PCGuy_5960]

14 minutes ago, cj09beira said:

Because x399 is coming later they will have time to resolve this issues.

LOL, do you think that designing a heatsink that can *actually* cool a VRM takes that long? 

[Curufinwe_wins]

7 minutes ago, Hunter259 said:

That is true but there is still heat added onto the board that on X299 is on the CPU. The ThreadRipper board is going to need a better cooling solution.

Completely Agreed.

 

 

With all due respect (to the intel bashers).... this looks 100% to be motherboard manufacturers cheaping out.... Look at the difference in the heatsink height and size between the x99 Deluxe and the x299 Deluxe.

 

The only part of this that would be on Intel is the rush, but even then... it would have been better just to reuse the old power delivery then over this shit.

 

Also here is the x99 Sabertooth compared to the x299 TUF... You can't honestly tell me anyone other than Asus (and other mobo makers) are to blame for that...

 

 

[Hunter259]

Just now, Curufinwe_wins said:

-snip-

With all due respect... how does that change what I said? Like at all? I said the cooling solutions need to be better. You say they are cheaping out. That in no way changes anything I said.

[The Benjamins]

1 minute ago, PCGuy_5960 said:

LOL, do you think that designing a heatsink that can *actually* cool a VRM takes that long? 

It allows for validation and revisions, that is how product design works.

 

if a product has less testing/refinement time issues like heat may not be solved. were if they had extra time they would see in testing that it get hot and go and revise the heat sink design to allow for better cooling.

 

lets say a product is launched 1 month early with would shorten testing and revision time by a lot, which may have a product release with issues.

 

I worked for a rapid engineering company and testing/tuning takes time and if shorten, corners will be cut or things will be missed. 

[Curufinwe_wins]

1 minute ago, Hunter259 said:

With all due respect... how does that change what I said? Like at all? I said the cooling solutions need to be better. You say they are cheaping out. That in no way changes anything I said.

Sorry, I was agreeing with you. will edit to make that apparent.

[cj09beira]

7 minutes ago, PCGuy_5960 said:

LOL, do you think that designing a heatsink that can *actually* cool a VRM takes that long? 

It doesnt take long to design it and i never said it would.

But with this happening now after release. With treadripper there is a good chance it will be good enough since day one

[Curufinwe_wins]

3 minutes ago, The Benjamins said:

It allows for validation and revisions, that is how product design works.

 

if a product has less testing/refinement time issues like heat may not be solved. were if they had extra time they would see in testing that it get hot and go and revise the heat sink design to allow for better cooling.

 

lets say a product is launched 1 month early with would shorten testing and revision time by a lot, which may have a product release with issues.

 

I worked for a rapid engineering company and testing/tuning takes time and if shorten, corners will be cut or things will be missed. 

Here is the thing though... they would have been better off just reusing the exact designs from last generation for heatsinks.... They already worked... The tolerances physically are the same (power delivery parts don't change much these days). And the x299 versions of the same boards are just FAR INFERIOR to anything they put out before them on the HEDT platform, which is inexcusable.

[rockon5622]

Looks like waterblocks for motherboards will be more of a thing lol. Better get wallet ready.

[done12many2]

52 minutes ago, Curufinwe_wins said:

Here is the thing though... they would have been better off just reusing the exact designs from last generation for heatsinks.... They already worked... The tolerances physically are the same (power delivery parts don't change much these days). And the x299 versions of the same boards are just FAR INFERIOR to anything they put out before them on the HEDT platform, which is inexcusable.

I'm using a Gigabyte x299 Aorus Gaming 7 to run my 7900x, which uses the same type dual heatsink with a heatpipe on the VRM as my Asus x99 Deluxe.  

 

Guess what?  No VRM temp issues at all.  No VRM fan installed or fans focused on the VRM. 

 

This is nothing more than an issue with cheaper boards that seem more available with x299.

[PCGuy_5960]

1 minute ago, done12many2 said:

This is nothing more than an issue with cheaper boards that seem more available with x299.

Yeah, which kinda sucks  

 

Idk, do you think that the AORUS X299 Gaming 3 or the ASUS X299-A will be able to handle an overclocked 7800X without overheating?

[pas008]

13 hours ago, Taf the Ghost said:

Well, Intel has paid a few billion USD in fines over their anti-competitive practices. They were able to "buy" AMD out of the mobile market by making it so no OEMs would make them, I believe it was.

X299 itself was within the launch window, but it was also a 10c max design until fairly recently.  Now it's an 18c max and it's very possible the Motherboard manufacturers simply didn't have time (or processors in house to test) to make all of the adjustments needed. I expect we'll be seeing a lot of v2 boards, especially if the rumor about power draw issues with the over 12c parts is true.

 

5 hours ago, SpaceGhostC2C said:

One of your links clearly says Q3.

Another link clearly shows Skylake-X as "6 to 10 cores"

 

Moving from Q3 to Q2, while adding models all the way to 18 cores right at launch date announcement doesn't really sound "as planned" to me.

not sure about you guys but considering the present and went those articles were written seems not rushed to me

and support for more cores doesnt constitute a flaw, how many times have you seen this fixed with a bios update x58/etc lol

 

obviously many companies werent paying attention considering the specs and leaks have showed actually what is now present

sounds like people bitching about their job and doing work becuz there is so many platforms and products right now

 

edited because 4th july weekend and spelling/drinking errors

which leads me to point another thing is if it is a vrm/etc issue why arent these manufacturers building these board to handle anything dropped(being the platform) in them considering many years of making them but we are on the snowflake generation now

 

[done12many2]

3 minutes ago, PCGuy_5960 said:

Yeah, which kinda sucks  

 

Idk, do you think that the AORUS X299 Gaming 3 or the ASUS X299-A will be able to handle an overclocked 7800X without overheating?

 

I think either will be fine, but once you start pushing high clocks on ANY 8 core or higher, they will start to draw some current.  With that said, you may need to actively cool the power delivery with a small fan if you start pushing high clocks.  

 

You could just avoid that altogether by getting a mid to high level board, but you'd have to make that call.  Also, Gigabytes BIOS is far from my favorite.  Asus really shines there.

 

There's too many overclocking and high core count experts in this thread to only listen to me so take in more opinions.  

[PCGuy_5960]

3 minutes ago, done12many2 said:

I think either will be fine, but once you start pushing high clocks on ANY 8 core or higher, they will start to draw some current.  With that said, you may need to actively cool the power delivery with a small fan if you start pushing high clocks.  

 

You could just avoid that altogether by getting a mid to high level board, but you'd have to make that call.  Also, Gigabytes BIOS is far from my favorite.  Asus really shines there.

The 7800X is a 6 core, so I think it should be fine   Although before I decide, I will wait for Coffeelake, so I'll see. If I go X299 I will probably buy an X299-A TBH and if the VRM runs hot, I'll just add a fan.

[MageTank]

Just now, done12many2 said:

 

I think either will be fine, but once you start pushing high clocks on ANY 8 core or higher, they will start to draw some current.  With that said, you may need to actively cool the power delivery with a small fan if you start pushing high clocks.  

 

You could just avoid that altogether by getting a mid to high level board, but you'd have to make that call.  Also, Gigabytes BIOS is far from my favorite.  Asus really shines there.

 

There's too many overclocking and high core count experts in this thread to only listen to me so take in more opinions.  

Honestly, even on higher end boards, I'd want active cooling. Gone are the days where you can have your cake and eat it too. If you want a high performance platform, time to throw away those gimmick aesthetic cases and buy something with real airflow (or a test rack with a fan hanging on top of ram/VRM's). 

 

I don't know if anyone else brought this up (i skipped like 10 pages of this thread) but his issue with VRM thermals seems questionable to some extent. Yes, some were throttling at their temperatures, but the temperatures were still below dangerous levels. I am not defending the manufacturers here, but I've certainly seen higher. Mitigate the thermal issues with active cooling, and you will be fine. The issue with the plug connectors is what worries me the most. While Jonnyguru came out on that video in the comments providing his thoughts (and how the cables themselves are not hot) the point of failure in that situation is almost always the plug connectors themselves. That much stress on a single 8pin EPS is simply madness, and they should have taken that into consideration. Plenty of X99 boards that provided more than just a single 8pin EPS, so X299 shouldn't have even considered a single 8pin.

 

As for everyone trying to let Intel walk away clean from this, ask yourself this. Which is more likely, intel giving out the proper specs, and several motherboard vendors got it wrong all to the same degree (weak VRM cooling and power delivery)? Or, Intel forgetting to provide the exact specs, and mobo manufacturers just winging it based off previous designs? I want to give Intel the benefit of the doubt, but it just seems unlikely that several board partners all made the exact same mistake, and that just be a coincidence. Could be another issue of these boards supplying incorrect amounts of voltages, like we saw on Z270, but that will take some time to figure out, as rapid bios updates are expected with a new launch.

 

Regardless, this launch has been pretty chaotic. To any of you planning on buying this platform, I'd advise you do as much research as possible to make sure that you get a decent board, and that it's features can be used by your CPU properly. Make sure you do not cheap out on the cooling solution, and for the love of all that is holy, do NOT throw AVX at it while overclocked until after you've seen what AVX does on stock clocks. Baby steps is key to avoiding that blue smoke, lol.

[VagabondWraith]

The VRM thing is blown wayyyy out of fucking proportion. People just parrot shit they see online.

[Hunter259]

25 minutes ago, done12many2 said:

I'm using a Gigabyte x299 Aorus Gaming 7 to run my 7900x, which uses the same type dual heatsink with a heatpipe on the VRM as my Asus x99 Deluxe.  

 

Guess what?  No VRM temp issues at all.  No VRM fan installed or fans focused on the VRM. 

 

This is nothing more than an issue with cheaper boards that seem more available with x299.

It's no excuse to have a shit cooling design on a $250+ motherboard....

[PCGuy_5960]

1 minute ago, VagabondWraith said:

The VRM thing is blown wayyyy out of fucking proportion. People just parrot shit they see online.

But, but Linus said that X299 is a failure! /s

[VagabondWraith]

1 minute ago, PCGuy_5960 said:

But, but Linus said that X299 is a failure! /s

That's just like, his opinion man. (Which I don't think much of.)

[PCGuy_5960]

Just now, VagabondWraith said:

That's just like, his opinion man. (Which I don't think much of.)

[done12many2]

14 minutes ago, Hunter259 said:

It's no excuse to have a shit cooling design on a $250+ motherboard....

 

A $250 HEDT board and a $250 mainstream board are two separate things. Sockets, extra trace routing for memory and other stuff take a lot out of the budget for HEDT boards, which results in a $250 board with shortcuts taken.  Remove all of those extra costs and you can increase costs spent on VRM cooling.  Or you could just make a board with all of the features and great VRM cooling and charge more money. 

 

Either way, your wrong because people of today expect it all for next to nothing or they will talk about you on LTT.  

 

12 minutes ago, VagabondWraith said:

That's just like, his opinion man. (Which I don't think much of.)

 

I watch Linus for the entertainment.  Not really to learn anything.

 

So I'm starting to figure out the BIOS on this Gigabyte board.  The IMCs on these Skylake-X chips seem pretty strong compared to previous x99 chips.  3600 c16 XMP was as simple as enable, save and reboot.  

 

Anyways, here was a run at just over 4.7 GHz.  I'm fairly confident I can break 2700cb before I end up delidding, but we'll see.  Single thread performance is fantastic on these things. 

 

 

[MageTank]

6 minutes ago, done12many2 said:

 

A $250 HEDT board and a $250 mainstream board are two separate things. Sockets, extra trace routing for memory and other stuff take a lot out of the budget for HEDT boards, which results in a $250 board with shortcuts taken.  Remove all of those extra costs and you can increase costs spent on VRM cooling.  Or you could just make a board with all of the features and great VRM cooling and charge more money. 

 

Either way, your wrong because people of today expect it all for next to nothing or they will talk about you on LTT.  

All the more reason to not trust the manufacturers to keep your components cool, lol. I need not explain this to you, as your loop is already beyond what I would consider "practical enough", but some people honestly expect to get away with using cheap AIO's on these CPU's and still hit high clock speeds. These people are in for a world of hurt, lol. The same goes with the cooling on your board. Always expect it to not be enough, and be prepared to remedy it if need be. 

 

There is a reason I own a ton of 40mm delta fans and a box of copper shims/ramsinks  

[done12many2]

1 minute ago, MageTank said:

All the more reason to not trust the manufacturers to keep your components cool, lol. I need not explain this to you, as your loop is already beyond what I would consider "practical enough", but some people honestly expect to get away with using cheap AIO's on these CPU's and still hit high clock speeds. These people are in for a world of hurt, lol. The same goes with the cooling on your board. Always expect it to not be enough, and be prepared to remedy it if need be. 

 

There is a reason I own a ton of 40mm delta fans and a box of copper shims/ramsinks  

 

My answer to any genius who says that a Skylake-X 10 core is really hot at high clock speeds.  No shit.   

 

When I read half the comments I see posted in threads like these I realize how few people are actually speaking from experience.  Hey, but they can theorize shit all day long!

[SpaceGhostC2C]

38 minutes ago, pas008 said:

 

and support for more cores doesnt constitute a flaw

You are missing the point: no one is saying support for more cores is a flaw. What we are saying is that it is harder for motherboard manufacturers to support cores they don't know will be there.

38 minutes ago, pas008 said:

 

obviously many companies werent paying attention considering the specs and leaks have showed actually what is now present

Are you not even reading your own leaks? The leaked specs were not what we are seeing: first 10 moved to 12, then Intel announced 18. Motherboard manufacturers have not yet seen these chips or even the specs. So no, X299 wasn't rushed in itself, i.e., they were working on some of it for a while. But they were working towards supporting a lower number of CPU models.

 

38 minutes ago, pas008 said:

which leads me to point another thing is if it is a vrm/etc issue why arent these manufacturers building these board to handle anything dropped(being the platform)

I have some ideas about it, but I already wrote them in previous posts, so I won't repeat them. Feel free to track back and read them  

 

38 minutes ago, pas008 said:

in them considering many years of making them but we are on the snowflake generation now

 

Umm... quit alcohol while you still have a liver.

[MageTank]

6 minutes ago, done12many2 said:

 

My answer to any genius who says that a Skylake-X 10 core is really hot at high clock speeds.  No shit.   

 

When I read half the comments I see posted in threads like these I realize how few people are actually speaking from experience.  Hey, but they can theorize shit all day long!

Oh, you already know my standards, so you already know I can't handle X299's heat, lol. At 5ghz on this CPU you sent me, I already enter the "toasty zone" under 48k FFT AVX2 load, which is around 70C. Now double the core count, and you have yourself a recipe for real heat, lol. People seriously expected 8 cores at 5ghz to be perfectly achievable on an AIO, and it makes me question their sanity. These CPU's are nowhere near as easy to delid as your standard 115x chips. We are dealing with a raised substrate on top of a ton of caps that are at risk of being covered in Intel's glue, which end up getting ripped completely off the substrate if done incorrectly. We also know 5ghz isn't easy to achieve without a delid on the quad cores, so expecting to get anywhere near that on the 8c+ SKU's on an AIO, it's just madness in it's purest form.

 

Luckily with X299, people are starting to understand there is more to it than just keeping the CPU cool. The VRM is going to matter. The thickness of your PSU cables are going to matter. The plugs on the board itself, will matter. For the first time in a long time, the enthusiast platform will actually require enthusiast level cooling and experience to get the most out of it. I think that is what is scaring people the most, lol.