Fan Configurations - (because I was very bored today....)

[NoddyPirate]

I’m sure i saw a Linus video, where he used a prime Z490 with a 10900K or something and it was fine. I assume the VRM’s on that are somewhat similar

I think with my board the 5600X just sits so far below the capability of the VRM that it just isn't a problem. Also, the relatively low number of phases it's putting out probably keeps the temperatures down too. But a 10900K would worry me!! At the same time, your average VRM can handle an awful lot of heat. I was watching a GN video recently where they were doing a crazy OC to get the their VRM on a cheap board up towards 120 degrees C. Not recommended obviously, but it happily carried on!

I was playing with the VRM phase control the other day - via the DIGI+ VRM thing that ASUS have on their more basic boards. Upping the phase switching frequency seemed to give a slightly smaller spread of voltages to the CPU cores - which is as you'd expect - and would be better for more critical OC's too - but would also have upped the VRM temperature in the process. It'll be another days work to figure out how much of a temp hike it actually creates.

High spec boards have big heatsinks not just to handle high power CPU's but also to be able to deal with their own high phase count and high switching frequency - to create that better quality power supply they give you. But I think these days, even lower spec boards are probably being better equipped to deal with temperatures than in the past. I certainly seem to only have memories of chipset heatsinks a decade or two ago, but otherwise exposed VRM components - even if I didn't actually know what they were at the time....!

 

[Ash_]

If you want to be impressed by VRM’s... Msi X570 Tomahawk... it literally rivals the Hero and boards in that price range. It has crazy power draw potential, i think the board and cpu can draw 350W or something and it needs to sustain that.

 

[AgentCooper]

I think with my board the 5600X just sits so far below the capability of the VRM that it just isn't a problem. Also, the relatively low number of phases it's putting out probably keeps the temperatures down too. But a 10900K would worry me!! At the same time, your average VRM can handle an awful lot of heat. I was watching a GN video recently where they were doing a crazy OC to get the their VRM on a cheap board up towards 120 degrees C. Not recommended obviously, but it happily carried on!

I was playing with the VRM phase control the other day - via the DIGI+ VRM thing that ASUS have on their more basic boards. Upping the phase switching frequency seemed to give a slightly smaller spread of voltages to the CPU cores - which is as you'd expect - and would be better for more critical OC's too - but would also have upped the VRM temperature in the process. It'll be another days work to figure out how much of a temp hike it actually creates.

High spec boards have big heatsinks not just to handle high power CPU's but also to be able to deal with their own high phase count and high switching frequency - to create that better quality power supply they give you. But I think these days, even lower spec boards are probably being better equipped to deal with temperatures than in the past. I certainly seem to only have memories of chipset heatsinks a decade or two ago, but otherwise exposed VRM components - even if I didn't actually know what they were at the time....!

I’d love to give you a bucket full of chips and different boards to play around with, I’ve enjoyed following your recent experiments. When I win the lottery I’ll get you a lab coat and load of silicon to play with 👍

 

[NoddyPirate]

I’d love to give you a bucket full of chips and different boards to play around with, I’ve enjoyed following your recent experiments. When I win the lottery I’ll get you a lab coat and load of silicon to play with 👍

I'll hold you to that @AgentCooper - It would be fun!!!

I'm having a ball passing the time pondering all this stuff. I think ultimately I am quite amazed at how accessible it all is these days though. The Interweb obviously helps in a huge way of course to get a grasp of things - but it doesn't seem that long ago that Overclocking was deemed only for spotty teens with nothing better to do than blow up a few CPU's in their bedrooms. And UEFI was more likely just a mispronounced UFO.

It's fun and interesting to me. And it must be said that having an AMD system has been a big key to all of this. So much more accessible than Intel from all I can tell......never mind the heat!

 

[AgentCooper]

I'll hold you to that @AgentCooper - It would be fun!!!

I'm having a ball passing the time pondering all this stuff. I think ultimately I am quite amazed at how accessible it all is these days though. The Interweb obviously helps in a huge way of course to get a grasp of things - but it doesn't seem that long ago that Overclocking was deemed only for spotty teens with nothing better to do than blow up a few CPU's in their bedrooms. And UEFI was more likely just a mispronounced UFO.

It's fun and interesting to me. And it must be said that having an AMD system has been a big key to all of this. So much more accessible than Intel from all I can tell......never mind the heat!

Note to self: I’ll also treat you to an i9-11900K, a pair of heatproof gloves and some fire insurance.

 

[NoddyPirate]

If you want to be impressed by VRM’s... Msi X570 Tomahawk... it literally rivals the Hero and boards in that price range. It has crazy power draw potential, i think the board and cpu can draw 350W or something and it needs to sustain that.

Just reading up there - the Tomahawk seems to have a max CPU power rating of 500W. The Crosshair HERO VIII is 1,000W!!!

But would you believe that my lowly PRIME B550 PLUS has a max of 395W?

Ryzen 65W TDP chips draw up to 76W from the VRM max (used to be 88W it seems) in default stock state - and 105W TDP chips go up to 142W. So even my cheapy board can easily handle all that. The quality of the supply is the big difference at those higher numbers though.

You can override the Ryzen default power limits above with Ryzen Master or in BIOS - but I'd not like to test that 395W limit on my yoke personally - unless AgentC was funding the experiment of course - but I was very surprised by it. Even if it is all a bit moot as the other Motherboard limits for raw current draw in Amps will be reached long before the Wattage limit at typical voltages.

 

[NoddyPirate]

Note to self: I’ll also treat you to an i9-11900K, a pair of heatproof gloves and some fire insurance.

Excellent thinking! And could you stretch to an outdoor test station also? I've only recently painted some rooms and I'd hate to have to redecorate..... 🤔

 

[AgentCooper]

Excellent thinking! And could you stretch to an outdoor test station also? I've only recently painted some rooms and I'd hate to have to redecorate..... 🤔

No, I’m going to stipulate that you perform all tests in a 4x4 plywood room surrounded by those red barrels you see in video games that explode as soon as you look at them. It’ll add an element of jeopardy to proceedings.

No pain, no gain 😜

And (as per our recent conversation thread) a nickname will be forthcoming...

I predict ‘Mr Burns’

 

[Ash_]

Just reading up there - the Tomahawk seems to have a max CPU power rating of 500W. The Crosshair HERO VIII is 1,000W!!!

But would you believe that my lowly PRIME B550 PLUS has a max of 395W?

Ryzen 65W TDP chips draw up to 76W from the VRM max (used to be 88W it seems) in default stock state - and 105W TDP chips go up to 142W. So even my cheapy board can easily handle all that. The quality of the supply is the big difference at those higher numbers though.

You can override the Ryzen default power limits above with Ryzen Master or in BIOS - but I'd not like to test that 395W limit on my yoke personally - unless AgentC was funding the experiment of course - but I was very surprised by it. Even if it is all a bit moot as the other Motherboard limits for raw current draw in Amps will be reached long before the Wattage limit at typical voltages.

How MSI can get it so right in one and so wrong in the other amazes me.

 

[NoddyPirate]

No, I’m going to stipulate that you perform all tests in a 4x4 plywood room surrounded by those red barrels you see in video games that explode as soon as you look at them. It’ll add an element of jeopardy to proceedings.

No pain, no gain 😜

And (as per our recent conversation thread) a nickname will be forthcoming...

I predict ‘Mr Burns’

In my first real job - which you now know about - I was originally hired to drive a Hiace around the country with six 50 gallon drums of fuel tied down with rope in the back. We'd then use a DC pump which I'd hook up to the van battery to pump fuel into machines downroute. I could always do some Intel testing in the back of that van if you prefer?

Actually I used to smoke while driving that thing too! (You should have seen the face on the odd hitch-hiker I'd pick up when they slowly realised the factors involved.) Every decent jeopardy experiment worth it's salt needs an ignition source - so I'll add a fag - lit by really cheap matches also.....

 

[TonyCarter]

Looks like my Asus ROG Strix X570 was the right compromise for me then - same temps as the Crosshair VII Hero

 

[Ash_]

Looks like my Asus ROG Strix X570 was the right compromise for me then - same temps as the Crosshair VII Hero

I think the Hero and Godlike and those premium boards, can handle those loads for extended periods. I’d assume the Tomahawk and Strix would start to throttle after a few hours

 

[TonyCarter]

I think the Hero and Godlike and those premium boards, can handle those loads for extended periods. I’d assume the Tomahawk and Strix would start to throttle after a few hours

Luckily my gaming sessions only last 8-12 hours nowadays...I'm getting old!

 

[Ash_]

Luckily my gaming sessions only last 8-12 hours nowadays...I'm getting old!

I don’t think gaming would be enough to cause throttling anyway 🤣.